This paper was converted on www.awesomepapers.org from LaTeX by an anonymous user.
Want to know more? Visit the Converter page.

Bi-center conditions and bifurcation of limit cycles in a class of Z2Z_{2}-equivariant cubic switching systems with two nilpotent points

Ting Chen [email protected] Feng Li [email protected] Yun Tian [email protected] Pei Yu [email protected] School of Statistics and Mathematics, Guangdong University of Finance and Economics,
Guangzhou, 510320, China
College of Science, National University of Defense Technology, Changsha, 410073, China School of Mathematics and Statistics, Linyi University, Linyi, 276005, China Department of Mathematics, Shanghai Normal University, Shanghai, 200234, China Department of Mathematics, Western University, London, Ontario, N6A 5B7, Canada
Abstract

In this paper, we generalize the Poincaré-Lyapunov method for systems with linear type centers to study nilpotent centers in switching polynomial systems and use it to investigate the bi-center problem of planar Z2Z_{2}-equivariant cubic switching systems associated with two symmetric nilpotent singular points. With a properly designed perturbation, 6 explicit bi-center conditions for such polynomial systems are derived. Then, based on the 66 center conditions, by using Bogdanov-Takens bifurcation theory with general perturbations, we prove that there exist at least 2020 small-amplitude limit cycles around the nilpotent bi-center for a class of Z2Z_{2}-equivariant cubic switching systems. This is a new lower bound of cyclicity for such cubic polynomial systems, increased from 1212 to 2020.

keywords:
Nilpotent singular point; Z2Z_{2}-equivariant switching system; bi-center; focus; cusp; limit cycle
MSC:
34C07, 34C23

1 Introduction

In recent years, there have been extensive studies on the qualitative behaviours of nonsmooth dynamical systems, associated with the vector fields which are either discontinuous or nondifferentiable. The main reason of the increasing interest in these studies is that many practical systems contain nonlinearity and nonsmoothness in their equations and display rich complex dynamical phenomena, as observed in biology [40], nonlinear oscillation [7, 29], dry friction of mechanical engineering [5, 30], power electronics [3] and so on. In fact, fundamental mathematical theory for such nonsmooth systems was established several decades ago, see [1, 41].

Switching system is one important class of nonsmooth systems, which is divided into several regions with different definitions of smooth vector fields. In this paper, we consider the switching systems described in the form of

(x˙,y˙)={(f+(x,y),g+(x,y)),ifS(x,y)Σ+,(f(x,y),g(x,y)),ifS(x,y)Σ,(\dot{x},\dot{y})=\left\{\begin{aligned} \left(f^{+}(x,y),\,g^{+}(x,y)\right),\quad\text{if}\ \ S(x,y)\in\Sigma^{+},\\ \left(f^{-}(x,y),\,g^{-}(x,y)\right),\quad\text{if}\ \ S(x,y)\in\Sigma^{-},\end{aligned}\right. (1)

where f±(x,y)f^{\pm}(x,y), g±(x,y)g^{\pm}(x,y) are analytic functions, and the boundary Σ\Sigma between these two regions Σ±\Sigma^{\pm} defines the switching manifold. Although the vector field could be neither continuous nor differentiable on the boundary Σ\Sigma, we recall the Filippov convention [18] that the vector field may be crossing, sliding on or escaping from the boundary. Before presenting our main results, we give an overview on two main problems in the qualitative theory of planar differential vector fields, namely, the center problem and the cyclicity problem.

We recall that an isolated singular point is monodromy in a planar vector field if all nearby orbits rotate about this point. It is well known [2] that a monodromic singular point can only be a center or a focus. The center problem is to determine whether a monodromic critical point is a center or a focus, see [20]. But this problem becomes extremely complicated in the switching systems described by (1). To overcome the difficulty, the authors of [21] presented a useful approach for computing the Lyapunov constants of switching systems, which can be used to derive the linear type center conditions for an elementary singular point characterized by a pair of purely imaginary eigenvalues. This type of centers can be determined by vanishing all the Lyapunov constants. In [9, 10, 24] a complete classification with conditions is given to determine if a singular point is a linear type center for several classes of switching systems. Nevertheless, for a given particular class of polynomial systems, finding a complete solution of the center problem is still extremely difficult.

For a given family of differential systems, the Lyapunov constants needed to solve the center problem are also related to the so called cyclicity problem, i.e., finding the maximum number of limit cycles around a singular point. The investigation on bifurcation of limit cycles in switching systems started a half century ago. In [26], Han and Zhang presented an interesting example to show that limit cycles can bifurcate in switching linear systems from a pseudo-focus, i.e., a focus of either focus-focus, parabolic-focus or parabolic-parabolic type, which cannot occur in smooth linear systems. Later, Freire et al. [19], Huan and Yang [28] independently proved that switching linear systems can have at least 33 limit cycles. Bastos et al. [4] proved that the number of crossing limit cycles in switching linear systems is at least 77, which has two regions separated by a cubic curve. Tian and Yu [42] constructed a special class of switching Bautin systems to show the existence of 1010 limit cycles in the neighborhood of a center. Recently, by using the averaging method, Braun et al. [6] obtained a new lower bound with 12 limit cycles bifurcating in quadratic switching systems. On the other hand, by the so-called pseudo-Hopf bifurcation analysis, one more limit cycle can be generated by a proper perturbation. However, not many works have been focused on the studies of cubic switching systems. Yu et al. [46] proved that cubic switching systems can have at least 18 limit cycles around two elementary foci. Chen et al. [12] presented a method to study the cyclicity problem in the neighborhood of the origin and infinity for a class of cubic switching systems. The maximal number of limit cycles obtained so far for cubic switching systems by perturbing an elementary center is 2626, see [23, 38].

The aim of this paper is to study the center and cyclicity problems associated with nilpotent singular points in switching systems. Let us recall some known results about nilpotent singular points. An isolated singular point of a planar polynomial system is said to be a nilpotent singular point if both eigenvalues of the Jacobian matrix of the system evaluated at this singular point are zero but the Jacobian matrix is not identically null. The authors of [14, 22, 31, 44, 39, 45] developed some computationally efficient methods for studying the nilpotent center problem of smooth systems. García [20] presented a technique for bounding the maximal number of limit cycles bifurcating from a family of nilpotent singular points in some symmetric smooth polynomial systems. Recently, Li et al. [32] obtained the conditions on two nilpotent singular points to be centers in a class of cubic smooth systems, and proved that 44 limit cycles exist around each of the two nilpotent singular points, with 44 more limit cycles bifurcating from two elementary second-order foci which are generated by breaking the first-order nilpotent foci, leading to a total of 1212 limit cycles, which is up to now the best result for cubic systems with nilpotent singular points.

However, the center problem and bifurcation of small-amplitude limit cycles for nilpotent singular points in switching systems become more challenging than that in the case of elementary centers. Facing the challenging, we generalize the Poincaré-Lyapunov method to compute what will be called generalized Lyapunov constants for switching nilpotent systems, and apply the method to study the bi-center problem and the cyclicity problem for a class of Z2Z_{2}-equivariant cubic switching systems with two symmetric nilpotent singular points. Here, we say that a ZqZ_{q}-equivariant differential system, whose global phase portraits are invariant under a rotation of 2π/q2\pi/q (qZ+q\in\mathrm{Z^{+}}) radians around the origin, has a bi-center at the singular points e1e_{1} and e2e_{2} if both the two singular points are centers, which will be called bi-center problem. The study of ZqZ_{q}-equivariant polynomial systems is closely related to the well-known Hilbert’s 16th problem, for more details see [13, 33, 34, 36, 37].

Without loss of generality, the Z2Z_{2}-equivariant cubic switching systems, based on (1) which satisfies (with the switching manifold Σ\Sigma defined as the xx-axis)

f+(x,y)=f(x,y),g+(x,y)=g(x,y),f^{+}(-x,-y)=-f^{-}(x,y),\quad g^{+}(-x,-y)=-g^{-}(x,y),

can be written in the form of the differential equations,

(x˙y˙)={(a00+a10x+a01y+a20x2+a11xy+a02y2+a30x3+a21x2y+a12xy2+a03y3,b00+b10x+b01y+b20x2+b11xy+b02y2+b30x3+b21x2y+b12xy2+b03y3,),ify>0,(a00+a10x+a01ya20x2a11xya02y2+a30x3+a21x2y+a12xy2+a03y3,b00+b10x+b01yb20x2b11xyb02y2+b30x3+b21x2y+b12xy2+b03y3,),ify<0,\left(\begin{array}[]{cc}\dot{x}\\ \dot{y}\end{array}\right)=\left\{\begin{aligned} &\left(\begin{aligned} &a_{00}+a_{10}x+a_{01}y+a_{20}x^{2}+a_{11}xy+a_{02}y^{2}\\ &+a_{30}x^{3}+a_{21}x^{2}y+a_{12}xy^{2}+a_{03}y^{3},\\ &b_{00}+b_{10}x+b_{01}y+b_{20}x^{2}+b_{11}xy+b_{02}y^{2}\\ &+b_{30}x^{3}+b_{21}x^{2}y+b_{12}xy^{2}+b_{03}y^{3},\end{aligned}\right),&\text{if}\ \ y>0,\\ &\left(\begin{aligned} &-a_{00}+a_{10}x+a_{01}y-a_{20}x^{2}-a_{11}xy-a_{02}y^{2}\\ &+a_{30}x^{3}+a_{21}x^{2}y+a_{12}xy^{2}+a_{03}y^{3},\\ &-b_{00}+b_{10}x+b_{01}y-b_{20}x^{2}-b_{11}xy-b_{02}y^{2}\\ &+b_{30}x^{3}+b_{21}x^{2}y+b_{12}xy^{2}+b_{03}y^{3},\end{aligned}\right),&\text{if}\ \ y<0,\end{aligned}\right. (2)

where the xx-axis is the switching manifold and all parameters are real. As is known, the type of nilpotent singular points can generate much more rich dynamics than that of the elementary one, such that the center of system (2) in the switching manifold can be made up of two center-focus (a center or a focus), or one cusp and one center-focus, or two cusps. In this paper, we assume that system (2) has two nilpotent singular points located at (±1,0)(\pm 1,0), and derive the conditions which ensure the nilpotent singular points (±1,0)(\pm 1,0) of (2) to be bi-center for two cases of critical points, namely, the third-order critical point and the second-order critical point. Moreover, we apply general perturbations to prove the existence of at least 2020 small-amplitude limit cycles bifurcating from the nilpotent bi-center (±1,0)(\pm 1,0), two of them are obtained from a symmetric pseudo-Hopf bifurcation by adding a suitable additional perturbation term. This is a new lower bound (increased from 1212 to 2020) on the number of limit cycles bifurcating in such cubic switching systems associated with nilpotent singular points.

The rest of the paper is organized as follows. In Section 2, we will simplify system (2) for the convenience of analysis and present our main results. In Section 3, we present some formulas which are needed in Sections 4 and 5 for proving the two main theorems. Section 4 is devoted to derive 66 conditions for (±1,0)(\pm 1,0) of system (2) to be bi-center. Then, in Section 5 we use the 66 center conditions to construct perturbed systems of (2) to show the bifurcation of 2020 limit cycles from the nilpotent singular points (±1,0)(\pm 1,0). Finally, conclusion with discussion on future work is given in Section 6.

2 Simplification of system (2) and the main results

Assuming that system (2) has two singular points at (±1,0)(\pm 1,0) yields that

a00=a20,b00=b20,a10=a30,b10=b30.a_{00}=-a_{20},\quad b_{00}=-b_{20},\quad a_{10}=-a_{30},\quad b_{10}=-b_{30}. (3)

The Jacobian matrices of (2) evaluated at (±1,0)(\pm 1,0) are given by

𝒥±=(±2a20+2a30a01±a11+a21±2b20+2b30b01±b11+b21).\mathcal{J}^{\pm}=\left(\begin{array}[]{cc}\pm 2a_{20}+2a_{30}&a_{01}\pm a_{11}+a_{21}\\ \pm 2b_{20}+2b_{30}&b_{01}\pm b_{11}+b_{21}\end{array}\right). (4)

It follows from 𝒥+=𝒥\mathcal{J}^{+}=\mathcal{J}^{-} that

a20=a11=b20=b11=0.a_{20}=a_{11}=b_{20}=b_{11}=0. (5)

To have (±1,0)(\pm 1,0) being isolated nilpotent singular points, the necessary and sufficient conditions are Tr(𝒥±)=det(𝒥±)=0{\rm Tr}(\mathcal{J}^{\pm})=\det(\mathcal{J}^{\pm})=0, but J±J^{\pm} is not identically zero. It is easy to obtain that a30=b01+b21=0a_{30}=b_{01}+b_{21}=0, and b30(a01+a21)=0b_{30}(a_{01}+a_{21})=0 (but not b30=a01+a21=0b_{30}=a_{01}+a_{21}=0), which gives either b300,a01+a21=0b_{30}\neq 0,\,a_{01}+a_{21}=0 or b30=0,a01+a210b_{30}=0,\,a_{01}+a_{21}\neq 0, that is

𝒥±=(002b300)or𝒥±=(0a01+a2100).\mathcal{J}^{\pm}=\left(\begin{array}[]{cc}0&0\\ 2b_{30}&0\end{array}\right)\qquad\textrm{or}\quad\mathcal{J}^{\pm}=\left(\begin{array}[]{cc}0&a_{01}+a_{21}\\ 0&0\end{array}\right). (6)

However, when b30=0b_{30}=0 and a01+a210a_{01}+a_{21}\neq 0, system (2) becomes

(x˙y˙)={((a01+a02y+a21x2+a12xy+a03y2)y,(b21+b02y+b21x2+b12xy+b03y2)y,),ify>0,((a01a02y+a21x2+a12xy+a03y2)y,(b21b02y+b21x2+b12xy+b03y2)y,),ify<0.\left(\begin{array}[]{cc}\dot{x}\\ \dot{y}\end{array}\right)=\left\{\begin{aligned} &\left(\begin{aligned} &(a_{01}+a_{02}y+a_{21}x^{2}+a_{12}xy+a_{03}y^{2})y,\\ &(-b_{21}+b_{02}y+b_{21}x^{2}+b_{12}xy+b_{03}y^{2})y,\end{aligned}\right),&\text{if}\ \ y>0,\\ &\left(\begin{aligned} &(a_{01}-a_{02}y+a_{21}x^{2}+a_{12}xy+a_{03}y^{2})y,\\ &(-b_{21}-b_{02}y+b_{21}x^{2}+b_{12}xy+b_{03}y^{2})y,\end{aligned}\right),&\text{if}\ \ y<0.\end{aligned}\right. (7)

It is easy to note that the polynomial equations in (7) have a common factor yy, and so (±1,0)(\pm 1,0) are not isolated singular points. Thus, b300b_{30}\neq 0.

Further, to make (±1,0)(\pm 1,0) be isolated nilpotent singular points of system (2), we set

𝒥±=(0010),\mathcal{J}^{\pm}=\left(\begin{array}[]{cc}0&0\\ 1&0\end{array}\right), (8)

which leads to b30=12b_{30}=\frac{1}{2}. With the above results, system (7) is reduced to

(x˙y˙)={(a21y+a02y2+a21x2y+a12xy2+a03y3,x2b21y+b02y2+x32+b21x2y+b12xy2+b03y3,),ify>0,(a21ya02y2+a21x2y+a12xy2+a03y3,x2b21yb02y2+x32+b21x2y+b12xy2+b03y3,),ify<0.\left(\begin{array}[]{cc}\dot{x}\\ \dot{y}\end{array}\right)=\left\{\begin{aligned} &\left(\begin{aligned} &-a_{21}y+a_{02}y^{2}+a_{21}x^{2}y+a_{12}xy^{2}+a_{03}y^{3},\\ &-\frac{x}{2}-b_{21}y+b_{02}y^{2}+\frac{x^{3}}{2}+b_{21}x^{2}y+b_{12}xy^{2}+b_{03}y^{3},\end{aligned}\right),&\text{if}\ \ y>0,\\[4.30554pt] &\left(\begin{aligned} &-a_{21}y-a_{02}y^{2}+a_{21}x^{2}y+a_{12}xy^{2}+a_{03}y^{3},\\ &-\frac{x}{2}-b_{21}y-b_{02}y^{2}+\frac{x^{3}}{2}+b_{21}x^{2}y+b_{12}xy^{2}+b_{03}y^{3},\end{aligned}\right),&\text{if}\ \ y<0.\end{aligned}\right. (9)

For the sake of convenience, we call the system in (9) for y>0y>0 “the upper system” and that for y<0y<0 “the lower system”.

Now, we need to discuss the multiplicity of the nilpotent singular points (±1,0)(\pm 1,0) for the upper and the lower systems in (9). In fact, by the symmetry of system (9), we only need to analyze the singular point (1,0)(1,0). Introducing the transformation xx+1x\rightarrow x+1 into system (9) results in

(x˙y˙)={(2a21xy+(a02+a12)y2+a21x2y+a12xy2+a03y3=Ψ+(x,y),x+3x22+2b21xy+(b02+b12)y2+x32+b21x2y+b12xy2+b03y3=Φ+(x,y)+x,),ify>0,(2a21xy+(a12a02)y2+a21x2y+a12xy2+a03y3=Ψ(x,y),x+3x22+2b21xy+(b12b02)y2+x32+b21x2y+b12xy2+b03y3=Φ(x,y)+x,),ify<0,\left(\begin{array}[]{cc}\dot{x}\\ \dot{y}\end{array}\right)=\left\{\begin{aligned} &\left(\begin{aligned} &2a_{21}xy+(a_{02}+a_{12})y^{2}+a_{21}x^{2}y+a_{12}xy^{2}\\ &+a_{03}y^{3}\stackrel{{\scriptstyle\triangle}}{{=}}\Psi^{+}(x,y),\\ &x+\frac{3x^{2}}{2}+2b_{21}xy+(b_{02}+b_{12})y^{2}+\frac{x^{3}}{2}+b_{21}x^{2}y\\ &+b_{12}xy^{2}+b_{03}y^{3}\stackrel{{\scriptstyle\triangle}}{{=}}\Phi^{+}(x,y)+x,\end{aligned}\right),&\text{if}\ \ y>0,\\[4.30554pt] &\left(\begin{aligned} &2a_{21}xy+(a_{12}-a_{02})y^{2}+a_{21}x^{2}y+a_{12}xy^{2}\\ &+a_{03}y^{3}\stackrel{{\scriptstyle\triangle}}{{=}}\Psi^{-}(x,y),\\ &x+\frac{3x^{2}}{2}+2b_{21}xy+(b_{12}-b_{02})y^{2}+\frac{x^{3}}{2}+b_{21}x^{2}y\\ &+b_{12}xy^{2}+b_{03}y^{3}\stackrel{{\scriptstyle\triangle}}{{=}}\Phi^{-}(x,y)+x,\end{aligned}\right),&\text{if}\ \ y<0,\end{aligned}\right. (10)

and thus the singular point (1,0)(1,0) of system (9) is shifted to the origin of system (10). Assume that

x±=𝔣±(y)=k=2ck±ykx^{\pm}=\mathfrak{f}^{\pm}(y)=\sum^{\infty}_{k=2}c_{k}^{\pm}y^{k}

are the unique solutions of the implicit function equations Φ±(x,y)+x=0\Phi^{\pm}(x,y)+x=0. Further, we denote that

F±(y)=Ψ±(𝔣±(y),y)=k=2𝔣k±yk,\displaystyle F^{\pm}(y)=\Psi^{\pm}(\mathfrak{f}^{\pm}(y),y)=\sum^{\infty}_{k=2}\mathfrak{f}_{k}^{\pm}y^{k}, (11)
G±(y)=[Ψ±x+Φ±y](𝔣±(y),y)=k=1𝔤k±yk,\displaystyle G^{\pm}(y)=\bigg{[}\frac{\partial\Psi^{\pm}}{\partial x}+\frac{\partial\Phi^{\pm}}{\partial y}\bigg{]}_{(\mathfrak{f}^{\pm}(y),y)}=\sum^{\infty}_{k=1}\mathfrak{g}_{k}^{\pm}y^{k},

where

𝔤1±=2(a21±b02+b12),\displaystyle\mathfrak{g}_{1}^{\pm}=2(a_{21}\pm b_{02}+b_{12}), (12)
𝔣2±=±a02+a12,\displaystyle\mathfrak{f}_{2}^{\pm}=\pm a_{02}+a_{12},
𝔣3±=a032a21b022a21b12.\displaystyle\mathfrak{f}_{3}^{\pm}=a_{03}\mp 2a_{21}b_{02}-2a_{21}b_{12}.

For planar smooth systems, if kk is the smallest integer satisfying 𝔣k0\mathfrak{f}_{k}\neq 0, then the multiplicity of the nilpotent singular point is exactly kk, for more details see [32]. Thus, we can use Theorem 3.5 in [16] to determine the type of the nilpotent singular point (0,0)(0,0) of (10) for the smooth polynomial equations either in the upper system or in the lower system.

More precisely, when 𝔤n=0\mathfrak{g}_{n}=0 and 𝔣m0\mathfrak{f}_{m}\neq 0, we have that

{m=2k+1,{𝔣m<0,(0,0) is a center or focus,𝔣m>0,(0,0) is a saddle,m=2k,(0,0) is a cusp.\left\{\begin{aligned} &m=2k+1,~{}\left\{\begin{aligned} &\mathfrak{f}_{m}<0,\ \textrm{$(0,0)$ is a center or focus},\\ &\mathfrak{f}_{m}>0,\ \textrm{$(0,0)$ is a saddle},\end{aligned}\right.\\ &m=2k,\ \textrm{$(0,0)$ is a cusp}.\\ \end{aligned}\right. (13)

When 𝔤n0\mathfrak{g}_{n}\neq 0, 𝔣m0\mathfrak{f}_{m}\neq 0 and Δ=4(n+1)𝔣m+𝔤n2\Delta=4(n+1)\mathfrak{f}_{m}+\mathfrak{g}_{n}^{2}, we have that

{m=2k+1,{𝔣m>0,(0,0) is a saddle,𝔣m<0,{k>n,ork=nandΔ0,{nodd, (0,0) is a H-E,neven, (0,0) is a node,k<n,ork=nandΔ<0,(0,0) is a center or focus,m=2k,{k>n,(0,0) is a saddle-node,kn,(0,0) is a cusp,\left\{\begin{aligned} &m=2k+1,\ \left\{\begin{aligned} &\mathfrak{f}_{m}>0,\ \textrm{$(0,0)$ is a saddle},\\ &\mathfrak{f}_{m}<0,\ \left\{\begin{aligned} &k>n,\ \textrm{or}\ k=n\ \textrm{and}\ \Delta\geq 0,\ \left\{\begin{aligned} &n\ \textrm{odd, $(0,0)$ is a H-E},\\ &n\ \textrm{even, $(0,0)$ is a node},\end{aligned}\right.\\ &k<n,\ \textrm{or}\ k=n\ \textrm{and}\ \Delta<0,\ \textrm{$(0,0)$ is a center or focus},\\ \end{aligned}\right.\end{aligned}\right.\\ &m=2k,\ \left\{\begin{aligned} &k>n,\ \textrm{$(0,0)$ is a saddle-node},\\ &k\leq n,\ \textrm{$(0,0)$ is a cusp},\end{aligned}\right.\end{aligned}\right. (14)

where H-E denotes the local phase portrait with one singular point consisting of one hyperbolic sector and one elliptic sector.

Proposition 2.1.

The multiplicity of the nilpotent singular point (1,0)(1,0) of the upper system ((or (1,0)(-1,0) of the lower system)) of (9) is at most 6.

Proof.

Using (12) with 𝔣2+=𝔣3+=0\mathfrak{f}_{2}^{+}=\mathfrak{f}_{3}^{+}=0, we have

a02=a12,a03=2a21(b02+b12).a_{02}=-a_{12},\quad a_{03}=2a_{21}(b_{02}+b_{12}).

First, assume that a21=0a_{21}\!=\!0. Then, we obtain 𝔣4+=a12(b02+b12)\mathfrak{f}_{4}^{+}\!=\!-a_{12}(b_{02}+b_{12}). If a12=0a_{12}\!=\!0, we have Ψ+(x,y)=0\Psi^{+}(x,y)=0, and so (1,0)(1,0) is not an isolated singular point, implying that 𝔣4+0\mathfrak{f}_{4}^{+}\neq 0. If b02=b12b_{02}=-b_{12}, we have 𝔣5+=a12b03\mathfrak{f}_{5}^{+}=-a_{12}b_{03}. Letting b03=0b_{03}=0 yields Ψ+(x,y)=a12xy2\Psi^{+}(x,y)=a_{12}xy^{2}, and then Ψ+(x,y)\Psi^{+}(x,y) and x+Φ+(x,y)x+\Phi^{+}(x,y) have a common factor xx. Hence, 𝔣5+0\mathfrak{f}_{5}^{+}\neq 0 if (1,0)(1,0) is an isolated singular point.

Next, assume that a210a_{21}\neq 0. Then, we have 𝔣4+=(b02+b12)(4a21b21a12)2a21b03\mathfrak{f}_{4}^{+}=(b_{02}+b_{12})(4a_{21}b_{21}-a_{12})-2a_{21}b_{03}. Setting

b03=(b02+b12)(4a21b21a12)2a21,b_{03}=\frac{(b_{02}+b_{12})(4a_{21}b_{21}-a_{12})}{2a_{21}},

we have 𝔣4+=0\mathfrak{f}_{4}^{+}=0 and

𝔣5+=(b02+b12)(a122+4a212b02+4a12a21b21)2a21.\mathfrak{f}_{5}^{+}=-\frac{(b_{02}+b_{12})(-a_{12}^{2}+4a_{21}^{2}b_{02}+4a_{12}a_{21}b_{21})}{2a_{21}}.

If b02=b12b_{02}=-b_{12}, we have b03=0b_{03}=0. Further, we obtain that Ψ+(x,y)\Psi^{+}(x,y) and Φ+(x,y)+x\Phi^{+}(x,y)+x have a common factor xx, leading to 𝔣5+0\mathfrak{f}_{5}^{+}\neq 0. Otherwise, we assume that b02=a1224a12a21b214a212b_{02}=\frac{a_{12}^{2}-4a_{12}a_{21}b_{21}}{4a_{21}^{2}}, under which 𝔣5+=0\mathfrak{f}_{5}^{+}=0. Then, we have

𝔣6+=a12(a122+4a212b124a12a21b21)232a214.\mathfrak{f}_{6}^{+}=\frac{a_{12}(a_{12}^{2}+4a_{21}^{2}b_{12}-4a_{12}a_{21}b_{21})^{2}}{32a_{21}^{4}}.

If a12=0a_{12}=0, we obtain that Ψ+(x,y)\Psi^{+}(x,y) and Φ+(x,y)+x\Phi^{+}(x,y)+x have a common factor 2x+x2+2b12y22x+x^{2}+2b_{12}y^{2}. If b12=a122+4a12a21b214a212b_{12}=\frac{-a_{12}^{2}+4a_{12}a_{21}b_{21}}{4a_{21}^{2}}, we have Ψ+(x,y)=xy[a21(2+x)+a12y]\Psi^{+}(x,y)=xy[a_{21}(2+x)+a_{12}y]. Then Ψ+(x,y)\Psi^{+}(x,y) and Φ+(x,y)+x\Phi^{+}(x,y)+x have a common factor xx. Hence, (1,0)(1,0) is not an isolated singular point, implying that 𝔣6+0\mathfrak{f}_{6}^{+}\neq 0. ∎

Next, we derive the conditions for the nilpotent singular points (±1\pm 1,0) of the Z2Z_{2}-equivariant cubic switching system (9) to be bi-center, yielding the following result.

Theorem 2.2.

The nilpotent singular points (±1\pm 1,0) of system (9) become bi-center if one of the following conditions holds:

I:\displaystyle\mathrm{I}\!: a02+a12=a21+b12=a12+3b03=b02=b21=0,a12>0,a03+2a212<0;\displaystyle\ \,a_{02}+a_{12}=a_{21}+b_{12}=a_{12}+3b_{03}=b_{02}=b_{21}=0,\ a_{12}>0,\ \,a_{03}+2a_{21}^{2}<0; (15)
II:\displaystyle\mathrm{II}\!: a02=a12=b02=b03=b21=0, 2a03+(b12a21)2<0;\displaystyle\ \,a_{02}=a_{12}=b_{02}=b_{03}=b_{21}=0,\ 2a_{03}+(b_{12}-a_{21})^{2}<0;
III:\displaystyle\mathrm{III}\!: a02=a12=b02=a21+b12=3b03+2a21b21\displaystyle\ \,a_{02}=a_{12}=b_{02}=a_{21}+b_{12}=3b_{03}+2a_{21}b_{21}
=9(a03+2a212)2+8a21b212(3a03+2a212)=0,a03+2a212<0;\displaystyle\hskip 20.95781pt=9(a_{03}+2a_{21}^{2})^{2}+8a_{21}b_{21}^{2}(3a_{03}+2a_{21}^{2})=0,\ \,a_{03}+2a_{21}^{2}<0;
IV:\displaystyle\mathrm{IV}\!: a02=a12=b02=8a21+3b212=16a033b212(4b12+b212)\displaystyle\ \,a_{02}=a_{12}=b_{02}=8a_{21}+3b_{21}^{2}=16a_{03}-3b_{21}^{2}(4b_{12}+b_{21}^{2})
=8b03b21(8b12b212)=0,9+438b212<b12<9438b212;\displaystyle\hskip 20.95781pt=8b_{03}-b_{21}(8b_{12}-b_{21}^{2})=0,\ \,-\tfrac{9+4\sqrt{3}}{8}\,b_{21}^{2}<b_{12}<-\tfrac{9-4\sqrt{3}}{8}\,b_{21}^{2};
V:\displaystyle\mathrm{V}\!: a12=b02=b03=b21=0,a02<0;\displaystyle\ \,a_{12}=b_{02}=b_{03}=b_{21}=0,\ a_{02}<0;
VI:\displaystyle\mathrm{VI}\!: a21+b12=a12+3b03=b02=b21=0,eithera02+|a12|<0ora02=a12<0.\displaystyle\ \,a_{21}+b_{12}=a_{12}+3b_{03}=b_{02}=b_{21}=0,\ \,\text{either}\ \,a_{02}+|a_{12}|<0\ \,\text{or}\ \,a_{02}=a_{12}<0.

Moreover, to find the maximal number of limit cycles around the nilpotent singular points (±1\pm 1,0) of system (9), we construct perturbed Z2Z_{2}-equivariant switching systems using the 66 bi-center conditions and prove that system (9) has at least 2020 limit cycles bifurcating from the nilpotent singular points (±1\pm 1,0). More precisely, we have the following theorem.

Theorem 2.3.

Under each of the 66 nilpotent bi-center conditions in Theorem 2.2, the Z2Z_{2}-equivariant cubic switching system (9) has at least 1818 limit cycles bifurcating from the two symmetric nilpotent singular points (±1,0)(\pm 1,0) by using all εk\varepsilon^{k}-order cubic perturbation, and at least 2020 limit cycles by, in addition, applying a symmetric pseudo-Hopf bifurcation.

3 The generalized Poincaré-Lyapunov method

In this section, for the convenience of reading, we first briefly describe the Poincaré-Lyapunov method for determining the linear type center in switching polynomial systems divided by the xx-axis, and then generate this method to study switching systems associated with isolated nilpotent singular points. As described in [10], we consider the following switching system,

(x˙,y˙)={(δxλ+y+i+j=2naij+xiyj,λ+x+δy+i+j=2nbij+xiyj),ify>0,(δxλy+i+j=2naijxiyj,λx+δy+i+j=2nbijxiyj),ify<0,(\dot{x},\,\dot{y})=\left\{\begin{aligned} \Big{(}\delta x-\lambda^{+}y+\sum\limits_{i+j=2}^{n}a_{ij}^{+}x^{i}y^{j},\ \lambda^{+}x+\delta y+\sum\limits_{i+j=2}^{n}b_{ij}^{+}x^{i}y^{j}\Big{)},\ \ \text{if}\ \ y>0,\\[0.0pt] \Big{(}\delta x-\lambda^{-}y+\sum\limits_{i+j=2}^{n}a_{ij}^{-}x^{i}y^{j},\ \lambda^{-}x+\delta y+\sum\limits_{i+j=2}^{n}b_{ij}^{-}x^{i}y^{j}\Big{)},\ \ \text{if}\ \ y<0,\end{aligned}\right. (16)

where δ,aij±,bij±R\delta,a_{ij}^{\pm},b_{ij}^{\pm}\in R, λ±>0\lambda^{\pm}>0. Let Λ±=(δ,aij±,bij±,λ±)\Lambda^{\pm}=(\delta,a_{ij}^{\pm},b_{ij}^{\pm},\lambda^{\pm}) represent two parameter vectors. The origin is a common singular point in both upper and lower systems of (16). With the polar coordinate transformation: x=rcosθx=r\cos\theta and y=rsinθy=r\sin\theta, system (16) can be rewritten as

drdθ={δr+i+j=3n+1cij+(aij+,bij+)cosθisinθjri+j1λ++i+j=3n+1dij+(aij+,bij+)cosθisinθjri+j2,ifθ[0,π],δr+i+j=3n+1cij(aij,bij)cosθisinθjri+j1λ+i+j=3n+1dij(aij,bij)cosθisinθjri+j2,ifθ[π,2π],\frac{{\rm d}r}{{\rm d}\theta}=\left\{\begin{aligned} &\frac{\delta r+\sum\limits_{i+j=3}^{n+1}c_{ij}^{+}(a_{ij}^{+},b_{ij}^{+})\cos\theta^{i}\sin\theta^{j}r^{i+j-1}}{\lambda^{+}+\sum\limits_{i+j=3}^{n+1}d_{ij}^{+}(a_{ij}^{+},b_{ij}^{+})\cos\theta^{i}\sin\theta^{j}r^{i+j-2}},\quad\text{if}\ \ \theta\in[0,\pi],\\[4.30554pt] &\frac{\delta r+\sum\limits_{i+j=3}^{n+1}c_{ij}^{-}(a_{ij}^{-},b_{ij}^{-})\cos\theta^{i}\sin\theta^{j}r^{i+j-1}}{\lambda^{-}+\sum\limits_{i+j=3}^{n+1}d_{ij}^{-}(a_{ij}^{-},b_{ij}^{-})\cos\theta^{i}\sin\theta^{j}r^{i+j-2}},\quad\text{if}\ \ \theta\in[\pi,2\pi],\end{aligned}\right. (17)

where cij±(aij±,bij±)c_{ij}^{\pm}(a_{ij}^{\pm},b_{ij}^{\pm}) and dij±(aij±,bij±)d_{ij}^{\pm}(a_{ij}^{\pm},b_{ij}^{\pm}) are polynomials in the parameters aij±a_{ij}^{\pm} and bij±b_{ij}^{\pm}. We suppose that the solutions for the upper and the lower systems of (17) are respectively obtained as r+(ϱ,Λ+,θ)=k1vk+(Λ+,θ)ϱkr^{+}(\varrho,\Lambda^{+},\theta)=\sum\limits_{k\geq 1}v_{k}^{+}(\Lambda^{+},\theta)\varrho^{k} and r(ϱ,Λ,θ)=k1vk(Λ+,θ)ϱkr^{-}(\varrho,\Lambda^{-},\theta)=\sum\limits_{k\geq 1}v_{k}^{-}(\Lambda^{+},\theta)\varrho^{k} satisfying the initial condition r+(ϱ,Λ+,0)=r(ϱ,Λ,π)=ϱr^{+}(\varrho,\Lambda^{+},0)=r^{-}(\varrho,\Lambda^{-},\pi)=\varrho. Then, we denote by

Υ+(ϱ)=r+(ϱ,Λ+,π)=eπδλ+ϱ+k2vk+ϱk\Upsilon^{+}(\varrho)=r^{+}(\varrho,\Lambda^{+},\pi)=e^{\pi\frac{\delta}{\lambda^{+}}}\varrho+\sum\limits_{k\geq 2}v_{k}^{+}\varrho^{k}

and

Υ(ϱ)=r(ϱ,Λ,2π)=eπδλϱ+k2vkϱk,\Upsilon^{-}(\varrho)=r^{-}(\varrho,\Lambda^{-},2\pi)=e^{\pi\frac{\delta}{\lambda^{-}}}\varrho+\sum\limits_{k\geq 2}v_{k}^{-}\varrho^{k},

the upper half-return map Υ+(ϱ)\Upsilon^{+}(\varrho) and the lower half-return map Υ(ϱ)\Upsilon^{-}(\varrho), respectively, where vk±v_{k}^{\pm}’s are the coefficients in Taylor expansions. Following the procedure in [21], we obtain the displacement function,

d(ϱ)=Υ+(ϱ)(Υ)1(ϱ)=k1Vkϱk,\displaystyle d(\varrho)=\Upsilon^{+}(\varrho)-(\Upsilon^{-})^{-1}(\varrho)=\sum\limits_{k\geq 1}V_{k}\varrho^{k}, (18)

where VkV_{k} is called the kkth-order Lyapunov constant of system (16). The origin of system (16) is a center if and only if all the Lyapunov constants in (18) vanish, i.e., d(ϱ)0d(\varrho)\equiv 0 for 0<ϱ10<\varrho\ll 1. If there exists χ(Λ+,Λ)\chi_{*}\in(\Lambda^{+},\Lambda^{-}) such that V1(χ)=V2(χ)==Vk(χ)=0V_{1}(\chi_{*})=V_{2}(\chi_{*})=\cdots=V_{k}(\chi_{*})=0 and Vk+1(χ)0V_{k+1}(\chi_{*})\neq 0, then any perturbations to system (16) can yield at most kk limit cycles bifurcating from the origin. Based on Lemma 4 in [42], we give the sufficient conditions for proving the existence of limit cycles bifurcating from the origin of system (16).

Lemma 3.1 ([42]).

If there exists a critical point χ=(a1c,a2c,,akc)\chi_{*}=(a_{1c},a_{2c},\cdots,a_{kc}) such that Vi1(χ)=Vi2(χ)==Vik(χ)=0V_{i_{1}}(\chi_{*})=V_{i_{2}}(\chi_{*})=\cdots=V_{i_{k}}(\chi_{*})=0, Vik+1(χ)0V_{i_{k+1}}(\chi_{*})\neq 0, with 1=i1<i2<<ik1=i_{1}<i_{2}<\cdots<i_{k}, and

det[(Vi1,Vi2,,Vik)(a1c,a2c,,akc)(χ)]0,\begin{array}[]{l}{\rm{det}}\bigg{[}\displaystyle\frac{\partial(V_{i_{1}},V_{i_{2}},\cdots,V_{i_{k}})}{\partial(a_{1c},a_{2c},\cdots,a_{kc})}(\chi_{*})\bigg{]}\neq 0,\end{array} (19)

then appropriate small perturbations about χ=χ\chi=\chi_{*} lead to that system (16) has exactly kk limit cycles bifurcating from the origin.

It is worth mentioning that a switching polynomial system can have one more small-amplitude limit cycle when the sliding segment changes its stability by adding constant terms, which is called pesudo-Hopf bifurcation, see [15, 18]. With δ>0\delta>0 and bb\in\mathbb{R}, we assume that f±(x,y)f^{\pm}(x,y) and g+(x,y)g^{+}(x,y) are given in (16), but g(x,y)g^{-}(x,y) has the following form

g(x,y)=b+λx+δy+i+j=2nbijxiyj.g^{-}(x,y)=b+\lambda^{-}x+\delta y+\sum\limits_{i+j=2}^{n}b_{ij}^{-}x^{i}y^{j}. (20)

Then, the upper system of (16) still has a singular point at the origin, while the lower system of (16) has a singular point near the origin. System (16) would have a sliding segment on the switching manifold y=0y=0 (the xx-axis). For small enough bb, the new switching system exhibits a pseudo-Hopf bifurcation at b=0b=0, as shown in Figure 1, which can produce one more small-amplitude limit cycle, see more details in [8].

\begin{overpic}[width=588.1196pt,height=108.83061pt]{Fig1.pdf} \put(18.0,2.0){$b<0$} \put(44.0,2.0){$b=0$} \put(69.0,2.0){$b>0$} \end{overpic}
Figure 1: Illustration of the idea of pseudo-Hopf bifurcation.

In the above discussions, we have briefly described the Poincaré-Lyapunov method and the pesudo-Hopf bifurcation for switching polynomial systems with linear type centers. Now, we consider the switching polynomial system with a nilpotent critical point at the origin,

(x˙,y˙)={(F1+(x,y),x+F2+(x,y)),ify>0,(F1(x,y),x+F2(x,y)),ify<0,(\dot{x},\,\dot{y})=\left\{\begin{aligned} \left(F_{1}^{+}(x,y),\ x+F_{2}^{+}(x,y)\right),\ \ \text{if}\ \ y>0,\\ \left(F_{1}^{-}(x,y),\ x+F_{2}^{-}(x,y)\right),\ \ \text{if}\ \ y<0,\end{aligned}\right. (21)

where F1±(x,y)F_{1}^{\pm}(x,y) and F2±(x,y)F_{2}^{\pm}(x,y) are real analytic functions without constant and linear terms. We will show that the Poincaré-Lyapunov method can be generalized to analyze center problem and bifurcation of limit cycles for the nilpotent singular point.

We give the following example to illustrate our idea, as the normal form of the nilpotent differential system can be simplified into the system (for example, see [25]),

x˙=\displaystyle\dot{x}= y,\displaystyle\ y, (22)
y˙=\displaystyle\dot{y}= cixi[1+g(x)]+dixi1y[1+h(x)]+y2q(x,y),\displaystyle\ c_{i}x^{i}[1+g(x)]+d_{i}x^{i-1}y[1+h(x)]+y^{2}q(x,y),

where i2i\geq 2, g(x)g(x), h(x)h(x) and q(x,y)q(x,y) are analytic functions satisfying g(0)=h(0)=0g(0)=h(0)=0. For the sake of simplicity, we consider a codimension-2 Bogdanov-Takens (B-T) bifurcation of the Z2Z_{2} cubic normal form (22). Then, we obtain the following normal form with unfolding:

x˙=\displaystyle\dot{x}= y,\displaystyle\ y, (23)
y˙=\displaystyle\dot{y}= αx+βy+c3x3+d3x2y,\displaystyle\ \alpha x+\beta y+c_{3}x^{3}+d_{3}x^{2}y,

where the term αx+βy\alpha x+\beta y is called unfolding with sufficiently small parameters α\alpha and β\beta. Note that the origin of system (23) is a linear type singular point when α<0\alpha<0 and β=0\beta=0. When β=0\beta=0, α0\alpha\rightarrow 0^{-}, the linear type origin becomes a nilpotent singular point.

We remark that the linear type origin of system (23) is a center if and only if d3=0d_{3}=0, and that the nilpotent monodromic origin (when c3<0c_{3}<0) of system (23) is also a center when d3=0d_{3}=0. Hence, the main idea of our method is to perturb system (21) and to establish a relation between the unperturbed system and the perturbed system based on the B-T bifurcation theory, which will generate perturbed Lyapunov constants.

To achieve this, we consider the following perturbed system of (21),

(x˙,y˙)={(ε2y+F1+(x,y)+εG1+(x,y,ε),x+F2+(x,y)+εG2+(x,y,ε)),ify>0,(ε2y+F1(x,y)+εG1(x,y,ε),x+F2(x,y)+εG2(x,y,ε)),ify<0,(\dot{x},\,\dot{y})=\left\{\begin{aligned} \left(-\varepsilon^{2}y+F_{1}^{+}(x,y)+\varepsilon G_{1}^{+}(x,y,\varepsilon),\ x+F_{2}^{+}(x,y)+\varepsilon G_{2}^{+}(x,y,\varepsilon)\right),\ \ \text{if}\ \ y>0,\\ \left(-\varepsilon^{2}y+F_{1}^{-}(x,y)+\varepsilon G_{1}^{-}(x,y,\varepsilon),\ x+F_{2}^{-}(x,y)+\varepsilon G_{2}^{-}(x,y,\varepsilon)\right),\ \ \text{if}\ \ y<0,\end{aligned}\right. (24)

where ε2y-\,\varepsilon^{2}y is called unfolding with sufficiently small |ε|1|\varepsilon|\!\ll\!1, G1±(x,y,ε)G_{1}^{\pm}(x,y,\varepsilon) and G2±(x,y,ε)G_{2}^{\pm}(x,y,\varepsilon) are real polynomials.

It should be noted that the ε\varepsilon-perturbation terms in (24) are applied to the whole system, not restricted to the local behavior. We use the ε2\varepsilon^{2} rather than ε\varepsilon in the linear perturbation term to avoid the εi\varepsilon^{-i} (i>0i>0) and |ε|\sqrt{|\varepsilon|} terms generalted in the later transformed systems. Based on B-T bifurcation theory and the relation established above for the two systems (21) and (24), we know that a nilpotent center can be ε\varepsilon-approximated by a linear type center. More detailed discussions on this subject can be found in [11].

For this type system of (24), we can apply the Poincaré-Lyapunov method and compute the generalized Lyapunov constants Vk(ε)V_{k}(\varepsilon). As a matter of fact, we have the displacement function of system (24), given by

d(ϱ)=k1Vk(ε)ϱk,whereVi(ε)=j=1εjVij,i=1,2,,d(\varrho)=\sum_{k\geq 1}V_{k}(\varepsilon)\varrho^{k},\quad\textrm{where}\quad V_{i}(\varepsilon)=\sum_{j=1}^{\infty}\varepsilon^{j}V_{ij},\quad i=1,2,\cdots, (25)

in which VijV_{ij} denotes the iith εj\varepsilon^{j}-order Lyapunov constant. We can determine the center conditions for system (24) by vanishing the ε\varepsilon terms in these generalized Lyapunov constants, thus leading to a set of algebraic conditions which characterize the existence of a nilpotent center of system (21).

Further, we can use the above generalized Poincaré-Lyapunov method to study the bifurcation of limit cycles from the nilpotent center of the switching system (21). By B-T bifurcation theory, we add a linear perturbation term ε2y-\varepsilon^{2}y to system (21) such that the origin of this switching system becomes a linear-type center. Then, we compute the generalized Lyapunov constants of the perturbed system with additional all εk\varepsilon^{k}-order perturbation terms. Following the procedure in [17], we can obtain the bifurcation of limit cycles from the nilpotent center as many as possible. Finally, we obtain one more limit cycle in the switching system by considering the pseudo-Hopf bifurcation.

4 The proof of Theorem 2.2

Since the multiplicity of monodromic critical point for smooth nilpotent systems is 2k+132k+1\geq 3 (see [16]), we know that the smallest multiplicity for a singular point is 33 if it is a nilpotent focus or center. For the sake of convenience, we call the singular point with multiplicity 33 the 3rd-order critical point. However, the multiplicity of the monodromic singular point in switching systems can be 22, i.e., a 2nd-order critical point. Hence, we consider four cases in the following two subsections.

4.1 The 33rd-order critical point (1,0)(1,0) of the upper system in (9)

By using the conditions given in (13) and (14), we obtain that the singular point (1,0)(1,0) of the upper system of (9) is a 3rd-order nilpotent focus or center if and only if

𝔣2+=0,𝔣3+<0,Δ+=(𝔤1+)2+8𝔣3+<0,\mathfrak{f}_{2}^{+}=0,\quad\mathfrak{f}_{3}^{+}<0,\quad\Delta^{+}=(\mathfrak{g}_{1}^{+})^{2}+8\mathfrak{f}_{3}^{+}<0, (26)

namely,

a02=a12,2a03+(b02+b12a21)2<0.\displaystyle a_{02}=-a_{12},\quad 2a_{03}+(b_{02}+b_{12}-a_{21})^{2}<0. (27)

Then, we obtain

𝔣2=2a12,𝔣3=a03+2a21(b02b12),Δ=8a03+4(a21+b02b12)2.\displaystyle\mathfrak{f}_{2}^{-}=2a_{12},\quad\mathfrak{f}_{3}^{-}=a_{03}+2a_{21}(b_{02}-b_{12}),\quad\Delta^{-}=8a_{03}+4(a_{21}+b_{02}-b_{12})^{2}.

When 𝔣20\mathfrak{f}_{2}^{-}\neq 0, i.e., a120a_{12}\neq 0, the singular point (1,0)(1,0) of the lower system in (9) is a cusp. On the other hand, the singular point (1,0)(1,0) of the lower system in (9) is a center or a focus with multiplicity 33 if 𝔣2=0\mathfrak{f}_{2}^{-}\!=0 (i.e., a12=0a_{12}=0), 𝔣3<0\mathfrak{f}_{3}^{-}\!<\!0 and Δ<0\Delta^{-}\!<\!0. In order to have a monodromic singular point at (1,0)(1,0) of (9), it requires that there do not exist seperatrices which connect this singular point in the upper system or the lower system. With the aid of Maple, we present the following example to demonstrate a phase portrait of system (9), indicating that the lower system of (9) has two seperatrices connecting (1,0)(1,0) in the lower half Poincaré disc when a12<0a_{12}<0. Thus, we only need to consider a120a_{12}\geq 0.

Example 4.1.

The phase portrait of system (9) with a02=b12=1a_{02}=b_{12}=1, a21=a12=1a_{21}=a_{12}=-1, a03=4a_{03}=-4, b02=b21=0b_{02}=b_{21}=0 and b03=13b_{03}=\frac{1}{3}, as depicted in Figure 2, shows that the singular points (±1,0)(\pm 1,0) are two cusps.

Refer to caption
Figure 2: The phase portrait of system (9) with a02=b12=1a_{02}\!=\!b_{12}\!=\!1, a21=a12=1a_{21}\!=\!a_{12}\!=\!-1,a03=4a_{03}\!=\!-4, b02=b21=0b_{02}\!=\!b_{21}=0 and b03=13b_{03}\!=\!\frac{1}{3}, showing that the singular points (±1,0)(\pm 1,0) are two cusps.

To discuss the bi-center conditions for (±1,0)(\pm 1,0) of system (9), we assume that a02=a12a_{02}=-a_{12}, as given in (27). Then, we show how to apply the Poincaré-Lyapunov method to derive the bi-center conditions for (±1,0)(\pm 1,0) of system (9). By the symmetry of system (9), we only need to study the center conditions at the singular point (1,0)(1,0).

With the transformation xx+1x\rightarrow x+1 applied and perturbations added to system (9) with (27), we obtain the following perturbed system, can check from (10)

(x˙y˙)={(ε2y+2a21xy+a21x2y+a12xy2+a03y3+k=1εkG1k+(x,y),x+3x22+2b21xy+(b02+b12)y2+x32+b21x2y+b12xy2+b03y3+k=1εkG2k+(x,y),),ify>0,(ε2y+2a21xy+2a12y2+a21x2y+a12xy2+a03y3+k=1εkG1k(x,y),x+3x22+2b21xy+(b12b02)y2+x32+b21x2y+b12xy2+b03y3+k=1εkG2k(x,y),),ify<0,\left(\begin{array}[]{cc}\dot{x}\\ \dot{y}\end{array}\right)=\left\{\begin{aligned} &\left(\begin{aligned} &-\varepsilon^{2}y+2a_{21}xy+a_{21}x^{2}y+a_{12}xy^{2}+a_{03}y^{3}\\ &+\sum_{k=1}\varepsilon^{k}G_{1k}^{+}(x,y),\\ &x+\frac{3x^{2}}{2}+2b_{21}xy+(b_{02}+b_{12})y^{2}+\frac{x^{3}}{2}\\ &+b_{21}x^{2}y+b_{12}xy^{2}+b_{03}y^{3}+\sum_{k=1}\varepsilon^{k}G_{2k}^{+}(x,y),\end{aligned}\right),&\text{if}\ \ y>0,\\[4.30554pt] &\left(\begin{aligned} &-\varepsilon^{2}y+2a_{21}xy+2a_{12}y^{2}+a_{21}x^{2}y+a_{12}xy^{2}+a_{03}y^{3}\\ &+\sum_{k=1}\varepsilon^{k}G_{1k}^{-}(x,y),\\ &x+\frac{3x^{2}}{2}+2b_{21}xy+(b_{12}-b_{02})y^{2}+\frac{x^{3}}{2}\\ &+b_{21}x^{2}y+b_{12}xy^{2}+b_{03}y^{3}+\sum_{k=1}\varepsilon^{k}G_{2k}^{-}(x,y),\end{aligned}\right),&\text{if}\ \ y<0,\end{aligned}\right. (28)

where

G1k±(x,y)=i+j=23pijk±xiyj,G2k±(x,y)=i+j=23qijk±xiyj,G_{1k}^{\pm}(x,y)=\sum_{i+j=2}^{3}p_{ijk}^{\pm}\,x^{i}y^{j},\quad G_{2k}^{\pm}(x,y)=\sum_{i+j=2}^{3}q_{ijk}^{\pm}\,x^{i}y^{j},

in which pijk±p_{ijk}^{\pm}, qijk±q_{ijk}^{\pm} are real parameters.

Note that system (28) should be invariant under the transformation xx1x\rightarrow x-1 (since (1,0)(1,0) and (1,0)(-1,0) are symmetric singular points of (9)), which yields

l20k±=l30k±=0,l11k=l11k+=2l21k+,l02k=l02k++2l12k+,l21k=l21k+,l12k=l12k+,l03k=l03k+,\begin{array}[]{llll}l_{20k}^{\pm}=l_{30k}^{\pm}=0,&l_{11k}^{-}=l_{11k}^{+}=2l_{21k}^{+},&l_{02k}^{-}=-l_{02k}^{+}+2l_{12k}^{+},\\[4.30554pt] l_{21k}^{-}=l_{21k}^{+},&l_{12k}^{-}=l_{12k}^{+},&l_{03k}^{-}=l_{03k}^{+},\end{array} (29)

where lijk±l_{ijk}^{\pm} represent pijk±p_{ijk}^{\pm} or qijk±q_{ijk}^{\pm}.

Since system (28) has a large number of perturbation parameters, which makes the calculation of the generalized Lyapuov constants very difficult, we need to reduce the number of parameters. Without loss generality, we assume that q21k+=0q_{21k}^{+}=0 because they are redundant for proving the necessity of the conditions I-VI. Further, we consider only the ε2\varepsilon^{2}-order perturbed terms, i.e., setting lij1+=lijk+=0l_{ij1}^{+}=l_{ijk}^{+}=0 (k>2k>2). The reason for why only choosing ε2\varepsilon^{2}-order perturbed terms will be given in Section 5.

For convenience, we let lij2+=lijl_{ij2}^{+}=l_{ij}. Further, introducing the transformation (x,y,t)(ε3x,ε2y,tε)(x,y,t)\rightarrow(\varepsilon^{3}x,\varepsilon^{2}y,\frac{t}{\varepsilon}) into system (28), we obtain

(x˙y˙)={(y+2(a21+p21ε2)εxy+p02ε2y2+(a21+p21ε2)ε4x2y+(a12+p12ε2)ε3xy2+(a03+p03ε2)ε2y3,x+32ε3x2+2b21ε2xy+(b02+b12+q02ε2)εy2+12ε6x3+b21ε5x2y+(b12+q12ε2)ε4xy2+(b03+q03ε2)ε3y3,),ify>0,(y+2(a21+p21ε2)εxy+(2a12p02ε2+2p12ε2)y2+(a21+p21ε2)ε4x2y+(a12+p12ε2)ε3xy2+(a03+p03ε2)ε2y3,x+32ε3x2+2b21ε2xy+(b12b02q02ε2+2q12ε2)εy2+12ε6x3+b21ε5x2y+(b12+q12ε2)ε4xy2+(b03+q03ε2)ε3y3,),ify<0.\left(\begin{array}[]{cc}\dot{x}\\ \dot{y}\end{array}\right)=\left\{\begin{aligned} &\left(\begin{aligned} &-y+2(a_{21}+p_{21}\varepsilon^{2})\varepsilon xy+p_{02}\varepsilon^{2}y^{2}+(a_{21}+p_{21}\varepsilon^{2})\varepsilon^{4}x^{2}y\\ &+(a_{12}+p_{12}\varepsilon^{2})\varepsilon^{3}xy^{2}+(a_{03}+p_{03}\varepsilon^{2})\varepsilon^{2}y^{3},\\[4.30554pt] &x+\tfrac{3}{2}\varepsilon^{3}x^{2}+2b_{21}\varepsilon^{2}xy+(b_{02}+b_{12}+q_{02}\varepsilon^{2})\varepsilon y^{2}+\tfrac{1}{2}\varepsilon^{6}x^{3}\\ &+b_{21}\varepsilon^{5}x^{2}y+(b_{12}+q_{12}\varepsilon^{2})\varepsilon^{4}xy^{2}+(b_{03}+q_{03}\varepsilon^{2})\varepsilon^{3}y^{3},\end{aligned}\right),&\text{if}\ \ y>0,\\[6.45831pt] &\left(\begin{aligned} &-y+2(a_{21}+p_{21}\varepsilon^{2})\varepsilon xy+(2a_{12}-p_{02}\varepsilon^{2}+2p_{12}\varepsilon^{2})y^{2}\\ &+(a_{21}+p_{21}\varepsilon^{2})\varepsilon^{4}x^{2}y+(a_{12}+p_{12}\varepsilon^{2})\varepsilon^{3}xy^{2}\\ &+(a_{03}+p_{03}\varepsilon^{2})\varepsilon^{2}y^{3},\\[4.30554pt] &x+\tfrac{3}{2}\varepsilon^{3}x^{2}+2b_{21}\varepsilon^{2}xy+(b_{12}-b_{02}-q_{02}\varepsilon^{2}+2q_{12}\varepsilon^{2})\varepsilon y^{2}\\ &+\tfrac{1}{2}\varepsilon^{6}x^{3}+b_{21}\varepsilon^{5}x^{2}y+(b_{12}+q_{12}\varepsilon^{2})\varepsilon^{4}xy^{2}\\ &+(b_{03}+q_{03}\varepsilon^{2})\varepsilon^{3}y^{3},\end{aligned}\right),&\text{if}\ \ y<0.\end{aligned}\right. (30)

To give a clear view of the proof, we first present a table below to show the flow of the proof.

Table 1: Outline of the proof for the center conditions I-IV.
Cases Conditions
(i) a02>0a_{02}>0 I
(ii) a02=0a_{02}=0 (ii-1) p12=b212p_{12}=-\tfrac{b_{21}}{2} II
(ii-2) b12=a21b_{12}=-a_{21} II
(ii-3) p02=p12p_{02}=p_{12} (ii-3-1) p12=1p_{12}=-1 (ii-3-1-1) b12=a21b_{12}=-a_{21} (ii-3-1-1-1) II, III
(ii-3-1-1-2)
(ii-3-1-2) b12a21b_{12}\neq-a_{21} (ii-3-1-2-1) II
(ii-3-1-2-2) II
(ii-3-1-2-3)
(ii-3-1-2-4) II, IV
(ii-3-2) q02=0q_{02}=0, 1+p2101+p_{21}\neq 0, (ii-3-2-1) b12=a21b_{12}=-a_{21}
               b21=13p12p21(2p211)b_{21}=\tfrac{1}{3}p_{12}p_{21}(2p_{21}\!-\!1) (ii-3-2-2) b12a21b_{12}\neq-a_{21}

The basic idea of the proof is setting the generalized Lyapunov constants zero to get a number of “necessary” center conditions, and excluding those which make certain order Lyapunov constant non-zero, and then later we prove that the remaining necessary conditions are also sufficient.

With the aid of a computer algebra system, we use the Poincaré-Lyapunov method to compute the generalized Lyapunov constants associated with the origin of system (30). The first two generalized Lyapunov constants are V1(ε)=0V_{1}(\varepsilon)=0, and

V2(ε)=\displaystyle V_{2}(\varepsilon)= 83ε[b02+(q02q12)ε2].\displaystyle\frac{8}{3}\varepsilon\,\big{[}b_{02}+(q_{02}-q_{12})\varepsilon^{2}\big{]}.

Setting the ε\varepsilon-order and ε3\varepsilon^{3}-order terms in V2(ε)V_{2}(\varepsilon) zero yield the necessary center conditions,

b02=q02q12=0.b_{02}=q_{02}-q_{12}=0.

Then, the 33rd generalized Lyapunov constant is given by

V3(ε)=\displaystyle V_{3}(\varepsilon)= π4ε{a12(a21+b12)+[2(a21+b12)p12+3b03+a12(1+2p21+2q02)2b12b21]ε2\displaystyle\frac{\pi}{4}\varepsilon\big{\{}a_{12}(a_{21}+b_{12})+\big{[}2(a_{21}+b_{12})p_{12}+3b_{03}+a_{12}(1+2p_{21}+2q_{02})-2b_{12}b_{21}\big{]}\varepsilon^{2}
+[3q032b21(1+q02)+p12(1+2p21+2q02)]ε4}.\displaystyle\qquad+\big{[}3q_{03}-2b_{21}(1+q_{02})+p_{12}(1+2p_{21}+2q_{02})\big{]}\varepsilon^{4}\big{\}}.

From the previous analysis for the multiplicity, we only consider two cases (i) a12>0a_{12}>0 and (ii) a12=0a_{12}=0.

  1. (i)

    Assume that a12>0a_{12}>0. Letting the ε\varepsilon-order, ε3\varepsilon^{3}-order and ε5\varepsilon^{5}-order terms in V3(ε)V_{3}(\varepsilon) equal zero we obtain the conditions,

    b12\displaystyle b_{12} =a21,b03=13[a12(1+2p21+2q02)+2a21b21],\displaystyle=-a_{21},\quad b_{03}=-\frac{1}{3}\big{[}a_{12}(1+2p_{21}+2q_{02})+2a_{21}b_{21}\big{]},
    q03\displaystyle q_{03} =13[2b21(1+q02)p12(1+2p21+2q02)].\displaystyle=\frac{1}{3}\big{[}2b_{21}(1+q_{02})-p_{12}(1+2p_{21}+2q_{02})\big{]}.

    Then, we have the 4th generalized Lyapunov constant, given by

    V4(ε)=\displaystyle V_{4}(\varepsilon)= 1645ε3[a12(p02p12)ε2]{4a12(p21+q02)+[3b21+2(b21+2p12)(p21+q02)]ε2}.\displaystyle-\frac{16}{45}\varepsilon^{3}\,\big{[}a_{12}-(p_{02}-p_{12})\varepsilon^{2}\big{]}\big{\{}4a_{12}(p_{21}+q_{02})+\big{[}3b_{21}+2(b_{21}+2p_{12})(p_{21}+q_{02})\big{]}\varepsilon^{2}\big{\}}.

    Thus, we obtain the conditions p21+q02=b21=0p_{21}+q_{02}=b_{21}=0 by setting the ε3\varepsilon^{3}-order and ε7\varepsilon^{7}-order terms in V4(ε)V_{4}(\varepsilon) zero. Therefore, the necessary center conditions are a21+b12=a12+3b03=b21=0a_{21}+b_{12}=a_{12}+3b_{03}=b_{21}=0, giving the condition I.

  1. (ii)

    Assume that a12=0a_{12}=0. By setting the ε3\varepsilon^{3}-order and ε5\varepsilon^{5}-order terms in V3(ε)V_{3}(\varepsilon) zero we obtain the conditions,

    b03=23[b12b21p12(a21+b12)],q03=13[2b21(1+q02)p12(1+2p21+2q02)].\displaystyle b_{03}=\frac{2}{3}\big{[}b_{12}b_{21}-p_{12}(a_{21}+b_{12})\big{]},\quad q_{03}=\frac{1}{3}\big{[}2b_{21}(1+q_{02})-p_{12}(1+2p_{21}+2q_{02})\big{]}.

    Then, we obtain the 44th generalized Lyapunov constant,

    V4(ε)=\displaystyle V_{4}(\varepsilon)= 1645ε5(p02p12){2(a21+b12)(b21+2p12)+[3b21+2(b21+2p12)(p21+q02)]ε2}.\displaystyle\frac{16}{45}\,\varepsilon^{5}(p_{02}-p_{12})\big{\{}2(a_{21}+b_{12})(b_{21}+2p_{12})+\big{[}3b_{21}+2(b_{21}+2p_{12})(p_{21}+q_{02})\big{]}\varepsilon^{2}\big{\}}.

    We have the following three subcases (ii-1) b21+2p12=0b_{21}+2p_{12}=0, (ii-2) a21+b12=0a_{21}+b_{12}=0 and (ii-3) p02p12=0p_{02}-p_{12}=0. Note that in the following analysis, when the condition in one case is satisfied, the conditions in other two cases may hold or not hold, depending upon the analysis for each case. In subcase (ii-3-1-1), both conditions p02p12=0p_{02}-p_{12}=0 and a21+b12=0a_{21}+b_{12}=0 are satisfied.

    (ii-1) Assume that p12=b212p_{12}=-\frac{b_{21}}{2}. Then, we have b21=0b_{21}=0 by setting V4(ε)=0V_{4}(\varepsilon)=0, yielding the condition II.

    (ii-2) Assume that b12=a21b_{12}=-a_{21}. Since the parameter q02q_{02} is redundant for the analysis of this subcase, without loss of generality, we let q02=0q_{02}=0. If p21=0p_{21}=0, we have b21=0b_{21}=0 by using V4(ε)=0V_{4}(\varepsilon)=0, which gives a special case of the condition II. Otherwise, if p210p_{21}\neq 0, we have p12=b21(3+2p21)4p21p_{12}=-\frac{b_{21}(3+2p_{21})}{4p_{21}} by setting ε7\varepsilon^{7}-order term of V4(ε)V_{4}(\varepsilon) zero. Then, the 5th generalized Lyapunov constant becomes

    V5(ε)=\displaystyle V_{5}(\varepsilon)= 1768b21πε7{16p21[20a21b2123(a03+2a2128a21b212)p21+2(3a0310a212)p212]\displaystyle\frac{1}{768}b_{21}\pi\,\varepsilon^{7}\big{\{}16p_{21}\big{[}20a_{21}b_{21}^{2}-3(a_{03}+2a_{21}^{2}-8a_{21}b_{21}^{2})p_{21}+2(3a_{03}-10a_{21}^{2})p_{21}^{2}\big{]}
    +[105b212+4b21(17b2190p02)p214(96a21+7b212+60b21p02+60p022+12p03)p212\displaystyle+\big{[}105b_{21}^{2}+4b_{21}(17b_{21}-90p_{02})p_{21}-4(96a_{21}\!+\!7b_{21}^{2}\!+\!60b_{21}p_{02}\!+\!60p_{02}^{2}\!+\!12p_{03})p_{21}^{2}
    +32(17a213b212)p213448a21p214]ε232p212(1+p21)(9+4p21+4p212)ε4}.\displaystyle\quad+32(17a_{21}-3b_{21}^{2})p_{21}^{3}-448a_{21}p_{21}^{4}\big{]}\varepsilon^{2}-32p_{21}^{2}(1+p_{21})\big{(}9+4p_{21}+4p_{21}^{2}\big{)}\varepsilon^{4}\big{\}}.

    Setting the ε11\varepsilon^{11}-order term in V5(ε)V_{5}(\varepsilon) zero we obtain p21=1p_{21}=-1. Further, letting V5(ε)=0V_{5}(\varepsilon)=0 we have

    a03=29(7a212+2a21b212),a21=196(29b21240b21p02+80p022+48p03).\displaystyle a_{03}=\frac{2}{9}(7a_{21}^{2}+2a_{21}b_{21}^{2}),\quad a_{21}=-\frac{1}{96}(29b_{21}^{2}-40b_{21}p_{02}+80p_{02}^{2}+48p_{03}).

    Letting the ε12\varepsilon^{12}-order term in V6(ε)V_{6}(\varepsilon) zero we obtain p03=0p_{03}=0. Then, the 6th generalized Lyapunov constant becomes

    V6(ε)=\displaystyle V_{6}(\varepsilon)= 118900b21(b214p02)ε9[3V6a+5V6bε2],\displaystyle-\frac{1}{18900}\,b_{21}(b_{21}-4p_{02})\,\varepsilon^{9}\big{[}3V_{6a}+5V_{6b}\varepsilon^{2}\big{]},

    where

    V6a=\displaystyle V_{6a}= (7b212+40b21p0280p022)(29b21240b21p02+80p022),\displaystyle\ (7b_{21}^{2}+40b_{21}p_{02}-80p_{02}^{2})(29b_{21}^{2}-40b_{21}p_{02}+80p_{02}^{2}),
    V6b=\displaystyle V_{6b}= 551b212+1544b21p023088p022.\displaystyle\ 551b_{21}^{2}+1544b_{21}p_{02}-3088p_{02}^{2}.

    By computing the resultant of V6aV_{6a} and V6bV_{6b} with respect to p02p_{02} we have

    Res[V6a,V6b,p02]=9011459133227925504b2180,(b210),\displaystyle{\rm Res}[V_{6a},V_{6b},p_{02}]=9011459133227925504\,b_{21}^{8}\neq 0,\quad(b_{21}\neq 0),

    which implies that V6aV_{6a} and V6bV_{6b} have no common roots. Hence, we obtain p02=b214p_{02}=\frac{b_{21}}{4} from V6(ε)=0V_{6}(\varepsilon)=0. Then, we obtain the 7th generalized Lyapunov constant,

    V7(ε)=\displaystyle V_{7}(\varepsilon)= 782944b215πε9(2b212+9ε2)(8b21+75ε2),\displaystyle-\frac{7}{82944}b_{21}^{5}\pi\,\varepsilon^{9}(2b_{21}^{2}+9\varepsilon^{2})(8b_{21}+75\varepsilon^{2}),

    which indicates that the ε9\varepsilon^{9}-order term in V7(ε)V_{7}(\varepsilon) is non-zero when b210b_{21}\neq 0.

    (ii-3) Assume that p02=p12p_{02}=p_{12}. Then, we have V4(ε)=0V_{4}(\varepsilon)=0 and obtain the 5th generalized Lyapunov constant,

    V5(ε)=π36ε5[3(a21+b12)V5a+V5bε2V5cε46V5dε6],\displaystyle V_{5}(\varepsilon)=\frac{\pi}{36}\varepsilon^{5}[3(a_{21}+b_{12})V_{5a}+V_{5b}\varepsilon^{2}-V_{5c}\varepsilon^{4}-6V_{5d}\varepsilon^{6}],

    where

    V5a=\displaystyle V_{5a}= 4a03b21+(5a03+4a2126a21b12)p12,\displaystyle\ 4a_{03}b_{21}+(5a_{03}+4a_{21}^{2}-6a_{21}b_{12})p_{12},
    V5b=\displaystyle V_{5b}= b21(9a0318a21b128b12b212)+3(a21+b12)(4b21+5p12)p03\displaystyle\ b_{21}(9a_{03}-18a_{21}b_{12}-8b_{12}b_{21}^{2})+3(a_{21}+b_{12})(4b_{21}+5p_{12})p_{03}
    +2[3(a21+b12)(a213b12)2(a214b12)b212]p12+10b21(a21+b12)p122\displaystyle+2[3(a_{21}+b_{12})(a_{21}-3b_{12})-2(a_{21}-4b_{12})b_{21}^{2}]p_{12}+10b_{21}(a_{21}+b_{12})p_{12}^{2}
    +3[4a03b21+(5a03+12a2124a21b126b122)p12]p21\displaystyle+3\big{[}4a_{03}b_{21}+(5a_{03}+12a_{21}^{2}-4a_{21}b_{12}-6b_{12}^{2})p_{12}\big{]}p_{21}
    +[6a21(a21+6b12)p123a03(4b21+5p12)]q02,\displaystyle+\big{[}6a_{21}(a_{21}+6b_{12})p_{12}-3a_{03}(4b_{21}+5p_{12})\big{]}q_{02},
    V5c=\displaystyle V_{5c}= 2b21(9a21+9b12+4b212)3b21(3+4p21)p03+6(a21+b122b212)p12\displaystyle\ 2b_{21}(9a_{21}+9b_{12}+4b_{21}^{2})-3b_{21}(3+4p_{21})p_{03}+6(a_{21}+b_{12}-2b_{21}^{2})p_{12}
    +18b12b21p2110(2b21+b21p21)p1226(6a21b12)p12p212(12a2112b12\displaystyle+18b_{12}b_{21}p_{21}-10(2b_{21}+b_{21}p_{21})p_{12}^{2}-6(6a_{21}-b_{12})p_{12}p_{21}^{2}-(12a_{21}-12b_{12}
    4b212+15p03)p12p21+18a21p12q022+[8b21316b212p122b21(6p03+5p122)\displaystyle-4b_{21}^{2}+15p_{03})p_{12}p_{21}+18a_{21}p_{12}q_{02}^{2}+\big{[}8b_{21}^{3}-16b_{21}^{2}p_{12}-2b_{21}(6p_{03}+5p_{12}^{2})
    3p12(5p0312b1212b12p21)+6a21(3b21+2p12+2p12p21)]q02,\displaystyle-3p_{12}(5p_{03}-12b_{12}-12b_{12}p_{21})+6a_{21}(3b_{21}+2p_{12}+2p_{12}p_{21})\big{]}q_{02},
    V5d=\displaystyle V_{5d}= (1+p21)[3b21+p12p21(12p21)+(3b21+p12+p12p21)q02+3p12q022].\displaystyle\ (1+p_{21})\big{[}3b_{21}+p_{12}p_{21}(1-2p_{21})+(3b_{21}+p_{12}+p_{12}p_{21})q_{02}+3p_{12}q_{02}^{2}\big{]}.

    Two subcases follow V5d=0V_{5d}=0: (ii-3-1) p21=1p_{21}=-1 and (ii-3-2)

    b21=13[p12p21(2p211)+(3b21+p12+p12p21)q02+3p12q022].b_{21}=\frac{1}{3}\big{[}p_{12}p_{21}(2p_{21}-1)+(3b_{21}+p_{12}+p_{12}p_{21})q_{02}+3p_{12}q_{02}^{2}\big{]}.

    (ii-3-1) When p21=1p_{21}=-1 we have V5d=0V_{5d}=0. In order to have that the ε5\varepsilon^{5}-order term in V5(ε)V_{5}(\varepsilon) equals zero, it needs b12=a21b_{12}=-a_{21} or V5a=0V_{5a}=0.

    (ii-3-1-1) Assume that b12=a21b_{12}=-a_{21}, we have a03+2a212<0a_{03}+2a_{21}^{2}<0. Then, we obtain the simplified V5bV_{5b},

    V5b=\displaystyle V_{5b}= b21[2a21(9a21+4b212)+3a03(4q021)]\displaystyle-b_{21}\big{[}2a_{21}(9a_{21}+4b_{21}^{2})+3a_{03}(4q_{02}-1)\big{]}
    +5[(3a03+6a212+4a21b212)3(a03+2a212)q02]p12.\displaystyle+5\big{[}(3a_{03}+6a_{21}^{2}+4a_{21}b_{21}^{2})-3(a_{03}+2a_{21}^{2})q_{02}\big{]}p_{12}.

    (ii-3-1-1-1) If

    q02=3a03+6a212+4a21b2123(a03+2a212),q_{02}=\frac{3a_{03}+6a_{21}^{2}+4a_{21}b_{21}^{2}}{3(a_{03}+2a_{21}^{2})},

    from V5b=0V_{5b}=0 we have b21=0b_{21}=0, which gives a special case of the condition II, or

    M1=9(a03+2a212)2+8a21b212(3a03+2a212)=0,\displaystyle M_{1}=9(a_{03}+2a_{21}^{2})^{2}+8a_{21}b_{21}^{2}(3a_{03}+2a_{21}^{2})=0,

    which does not contain the perturbed parameters, as expected, giving the condition III. Further there exists the free parameter p03p_{03} such that V5c=0V_{5c}=0.

    (ii-3-1-1-2) If

    q023a03+6a212+4a21b2123(a03+2a212),q_{02}\neq\frac{3a_{03}+6a_{21}^{2}+4a_{21}b_{21}^{2}}{3(a_{03}+2a_{21}^{2})},

    we have

    p12=b21[2a21(9a21+4b212)+3a03(4q021)]5(3a03+6a212+4a21b212)15(a03+2a212)q02\displaystyle p_{12}=\frac{b_{21}\big{[}2a_{21}(9a_{21}+4b_{21}^{2})+3a_{03}(4q_{02}-1)\big{]}}{5(3a_{03}+6a_{21}^{2}+4a_{21}b_{21}^{2})-15(a_{03}+2a_{21}^{2})q_{02}}

    from V5b=0V_{5b}=0. Letting

    M2=2a21(1q02)2b212(12q02)=0M_{2}=-2a_{21}(1-q_{02})^{2}-b_{21}^{2}(1-2q_{02})=0

    from V5c=0V_{5c}=0, which yields V5(ε)=V6(ε)=0V_{5}(\varepsilon)=V_{6}(\varepsilon)=0. Then, from the ε9\varepsilon^{9}-order term of V7(ε)V_{7}(\varepsilon),

    V7a=\displaystyle V_{7a}= b21M1=0\displaystyle\,b_{21}M_{1}=0

    which also giving the special case of the condition II and the condition III. In fact, from the ε11\varepsilon^{11}-order term V7bV_{7b} and ε13\varepsilon^{13}-order term V7cV_{7c} of V7(ε)V_{7}(\varepsilon), we have the resultants of M2M_{2}, V7bV_{7b}, and V7cV_{7c} with respect to q02q_{02}, respectively,

    Res[M2,V7b,V7c,q02]=[0,0],{\rm Res}[M_{2},V_{7b},V_{7c},q_{02}]=[0,0],

    i.e., there exist the free parameters such V7b=V7c=0V_{7b}=V_{7c}=0. Then we have V7(ε)=0V_{7}(\varepsilon)=0.

    We remark that since the parameter q02q_{02} is redundant for the following subcases, without loss of generality, we let q02=0q_{02}=0.

    (ii-3-1-2) Assume that b12a21b_{12}\neq-a_{21}. We solve the polynomial equations: V5a=V5b=V5c=0V_{5a}=V_{5b}=V_{5c}=0 to find other real solutions for parameters a21a_{21}, a03a_{03}, b12b_{12} and b21b_{21}.

    (ii-3-1-2-1) If a03=25a21(2a213b12)<0a_{03}\!=\!-\,\frac{2}{5}a_{21}(2a_{21}\!-\!3b_{12})\!<\!0, we have b21=0b_{21}\!=\!0, and either p12=0p_{12}\!=\!0 or 6a215p03=06a_{21}-5p_{03}\!=\!0 by using the equations V5a=V5b=V5c=0V_{5a}=V_{5b}=V_{5c}=0. The first solution is included in the condition II. For the second solution, if p03=6a215p_{03}=\frac{6a_{21}}{5} we obtain V6(ε)=0V_{6}(\varepsilon)=0 and the 7th generalized Lyapunov constant, given by

    V7(ε)=\displaystyle V_{7}(\varepsilon)= π180ε9a21p12(6a2135p122)[(a21+b12)2ε4].\displaystyle-\frac{\pi}{180}\varepsilon^{9}a_{21}p_{12}(6a_{21}-35p_{12}^{2})\big{[}(a_{21}+b_{12})^{2}-\varepsilon^{4}\big{]}.

    Setting V7(ε)=0V_{7}(\varepsilon)=0 yields a21=356p1220a_{21}=\frac{35}{6}p_{12}^{2}\neq 0, leading to V8(ε)=0V_{8}(\varepsilon)=0 and

    V9(ε)=98729p127πε13(6b12+35p1226ε2)(6b12+35p122+6ε2).\displaystyle V_{9}(\varepsilon)=-\frac{98}{729}p_{12}^{7}\pi\varepsilon^{13}(6b_{12}+35p_{12}^{2}-6\varepsilon^{2})(6b_{12}+35p_{12}^{2}+6\varepsilon^{2}).

    Hence, we obtain that the ε17\varepsilon^{17}-order term in V9(ε)V_{9}(\varepsilon) is not vanished.

    (ii-3-1-2-2) If a21=0a_{21}=0, by setting V5a=V5b=V5c=V7(ε)=0V_{5a}=V_{5b}=V_{5c}=V_{7}(\varepsilon)=0 we obtain b21=0b_{21}=0, which is included in the condition II.

    (ii-3-1-2-3) If 2a213b12=02a_{21}-3b_{12}=0, we obtain

    b12=32115b212,a21=48115b212,a03=307213225b214,p12=45b21,p03=56575b212,\displaystyle b_{12}=-\frac{32}{115}b_{21}^{2},\quad a_{21}=-\frac{48}{115}b_{21}^{2},\quad a_{03}=-\frac{3072}{13225}b_{21}^{4},\quad p_{12}=-\frac{4}{5}b_{21},\quad p_{03}=\frac{56}{575}b_{21}^{2},

    by solving V5a=V5b=V5c=V7(ε)=0V_{5a}=V_{5b}=V_{5c}=V_{7}(\varepsilon)=0. Further, we have V8(ε)=0V_{8}(\varepsilon)=0 and

    V9(ε)=322224π502839296875b217ε13(24576b21413984b212ε210051ε4)0ifb210.\displaystyle V_{9}(\varepsilon)=\frac{322224\pi}{502839296875}b_{21}^{7}\varepsilon^{13}(24576b_{21}^{4}-13984b_{21}^{2}\varepsilon^{2}-10051\varepsilon^{4})\neq 0\quad\textrm{if}\ b_{21}\neq 0.

    But if b212=23(19+2185)ε21536b_{21}^{2}=\frac{23(19+\sqrt{2185})\varepsilon^{2}}{1536}, then V9(ε)=0V_{9}(\varepsilon)=0.

    (ii-3-1-2-4) Assume that a21(2a213b12)(5a03+4a2126a21b12)0a_{21}(2a_{21}-3b_{12})(5a_{03}+4a_{21}^{2}-6a_{21}b_{12})\neq 0. If p12=0p_{12}=0, we have b21=0b_{21}=0 from V5a=0V_{5a}=0, which is a special case of the condition II. Suppose that p120p_{12}\neq 0. Then, by using V5a=V5b=0V_{5a}=V_{5b}=0, we have that

    b21=1a03p12(5a03+4a2126a21b12)0,\displaystyle b_{21}=-\frac{1}{a_{03}}\,p_{12}(5a_{03}+4a_{21}^{2}-6a_{21}b_{12})\neq 0,
    p03=148a033a21(a21+b12)(2a213b12)M3,\displaystyle p_{03}=\frac{1}{48a_{03}^{3}a_{21}(a_{21}+b_{12})(2a_{21}-3b_{12})}\,M_{3},

    where

    M3=\displaystyle M_{3}= 18a032[5a032+4a03a21(3a212b12)4a212b12(2a213b12)]\displaystyle\ 18a_{03}^{2}\big{[}5a_{03}^{2}+4a_{03}a_{21}(3a_{21}-2b_{12})-4a_{21}^{2}b_{12}(2a_{21}-3b_{12})\big{]}
    +p122(2a213b12)(a032a21b12)(5a03+4a2126a21b12)(15a03+4a2126a21b12).\displaystyle+p_{12}^{2}(2a_{21}-3b_{12})(a_{03}-2a_{21}b_{12})(5a_{03}+4a_{21}^{2}-6a_{21}b_{12})(15a_{03}+4a_{21}^{2}-6a_{21}b_{12}).

    Then, the polynomial V5cV_{5c} becomes

    V5c=(5a03+4a2126a21b12)p1264a033a21(a21+b12)(2a213b12)M4M5,\displaystyle V_{5c}=\frac{(5a_{03}+4a_{21}^{2}-6a_{21}b_{12})p_{12}}{64a_{03}^{3}a_{21}(a_{21}+b_{12})(2a_{21}-3b_{12})}M_{4}M_{5},

    where

    M4=\displaystyle M_{4}= 15a032+4a212[5a03+(2a213b12)(4a21+5b12)],\displaystyle\ 15a_{03}^{2}+4a_{21}^{2}\big{[}5a_{03}+(2a_{21}-3b_{12})(4a_{21}+5b_{12})\big{]},
    M5=\displaystyle M_{5}= 18a032+[15a03(2a213b12)+2a21(2a213b12)2]p122.\displaystyle\ 18a_{03}^{2}+\big{[}15a_{03}(2a_{21}-3b_{12})+2a_{21}(2a_{21}-3b_{12})^{2}\big{]}p_{12}^{2}.

    To have V5c=0V_{5c}=0, it requires M4M5=0M_{4}M_{5}=0 under which V6(ε)=0V_{6}(\varepsilon)=0, and the 7th generalized Lyapunov constant becomes

    V7(ε)=\displaystyle V_{7}(\varepsilon)= πε93538944a035a212(2a213b12)2(a21+b12)2\displaystyle\ \frac{\pi\varepsilon^{9}}{3538944a_{03}^{5}a_{21}^{2}(2a_{21}-3b_{12})^{2}(a_{21}+b_{12})^{2}}
    ×[1536a032a212(2a213b12)2(a21+b12)2(a032a21b12)V~7a\displaystyle\times\big{[}1536a_{03}^{2}a_{21}^{2}(2a_{21}-3b_{12})^{2}(a_{21}+b_{12})^{2}(a_{03}-2a_{21}b_{12})\widetilde{V}_{7a}
    +16a21(2a213b12)(a21+b12)(5a03+4a2126a21b12)V~7bε2V~7cε4],\displaystyle\quad\ +16a_{21}(2a_{21}-3b_{12})(a_{21}+b_{12})(5a_{03}+4a_{21}^{2}-6a_{21}b_{12})\widetilde{V}_{7b}\varepsilon^{2}-\widetilde{V}_{7c}\varepsilon^{4}\big{]},

    where

    V~7a=\displaystyle\widetilde{V}_{7a}= 630a034+15a033[8a21(8a2113b12)+(46a21129b12)p122]\displaystyle\ 630a_{03}^{4}+15a_{03}^{3}\big{[}8a_{21}(8a_{21}-13b_{12})+(46a_{21}-129b_{12})p_{12}^{2}\big{]}
    +2a032a21(2a213b12)[36a21(2a215b12)+25(22a2127b12)p122]\displaystyle+2a_{03}^{2}a_{21}(2a_{21}-3b_{12})\big{[}36a_{21}(2a_{21}-5b_{12})+25(22a_{21}-27b_{12})p_{12}^{2}\big{]}
    +140a03a212(2a213b12)3p122+8a213(2a215b12)(2a213b12)3p122,\displaystyle+140a_{03}a_{21}^{2}(2a_{21}-3b_{12})^{3}p_{12}^{2}+8a_{21}^{3}(2a_{21}-5b_{12})(2a_{21}-3b_{12})^{3}p_{12}^{2},
    V~7b=\displaystyle\widetilde{V}_{7b}= 5670a037+135a036[2368a212+6a21(986b1219p122)+b12(2792b12+411p122)]\displaystyle\ 5670a_{03}^{7}+135a_{03}^{6}\big{[}2368a_{21}^{2}+6a_{21}(986b_{12}-19p_{12}^{2})+b_{12}(2792b_{12}+411p_{12}^{2})\big{]}
    +18a035[4a212(6396a212+16316a21b12+3305b122)+10(2614a213+4917a212b12\displaystyle+18a_{03}^{5}\big{[}4a_{21}^{2}(6396a_{21}^{2}+16316a_{21}b_{12}+3305b_{12}^{2})+10(2614a_{21}^{3}+4917a_{21}^{2}b_{12}
    9407a21b1225235b123)p12225(46a21129b12)(2a213b12)p124]\displaystyle-9407a_{21}b_{12}^{2}-5235b_{12}^{3})p_{12}^{2}-25(46a_{21}-129b_{12})(2a_{21}-3b_{12})p_{12}^{4}\big{]}
    +12a034a21(2a213b12)[12a21(2696a213+8966a212b12+9067a21b122+3490b123)\displaystyle+12a_{03}^{4}a_{21}(2a_{21}-3b_{12})\big{[}12a_{21}(2696a_{21}^{3}+8966a_{21}^{2}b_{12}+9067a_{21}b_{12}^{2}+3490b_{12}^{3})
    +30(1122a213+3451a212b12540a21b122349b123)p1225(1376a2124878a21b12\displaystyle+30(1122a_{21}^{3}+3451a_{21}^{2}b_{12}-540a_{21}b_{12}^{2}-349b_{12}^{3})p_{12}^{2}-5(1376a_{21}^{2}-4878a_{21}b_{12}
    +5121b122)p124]+16a033a212(2a213b12)2p122[22882a213+78167a212b12\displaystyle+5121b_{12}^{2})p_{12}^{4}\big{]}+16a_{03}^{3}a_{21}^{2}(2a_{21}-3b_{12})^{2}p_{12}^{2}\big{[}22882a_{21}^{3}+78167a_{21}^{2}b_{12}
    +65395a21b122+26175b12315(180a212854a21b12+891b122)p122]\displaystyle+65395a_{21}b_{12}^{2}+26175b_{12}^{3}-15(180a_{21}^{2}-854a_{21}b_{12}+891b_{12}^{2})p_{12}^{2}\big{]}
    +16a032a213(2a213b12)3p122[2696a213+9182a212b12+8527a21b122+3490b123\displaystyle+16a_{03}^{2}a_{21}^{3}(2a_{21}-3b_{12})^{3}p_{12}^{2}\big{[}2696a_{21}^{3}+9182a_{21}^{2}b_{12}+8527a_{21}b_{12}^{2}+3490b_{12}^{3}
    30(16a212140a21b12+159b122)p122]+192a215(2a215b12)(2a213b12)5b12p124\displaystyle-30(16a_{21}^{2}-140a_{21}b_{12}+159b_{12}^{2})p_{12}^{2}\big{]}+192a_{21}^{5}(2a_{21}-5b_{12})(2a_{21}-3b_{12})^{5}b_{12}p_{12}^{4}
    96a03a214(2a213b12)4(4a21296a21b12+145b122)p124,\displaystyle-96a_{03}a_{21}^{4}(2a_{21}-3b_{12})^{4}(4a_{21}^{2}-96a_{21}b_{12}+145b_{12}^{2})p_{12}^{4},

    and

    V~7c=\displaystyle\widetilde{V}_{7c}= 4252500a039+56700a038[6a21(46a2119b12)+125(2a213b12)p122]\displaystyle\ 4252500a_{03}^{9}+56700a_{03}^{8}\big{[}6a_{21}(46a_{21}-19b_{12})+125(2a_{21}-3b_{12})p_{12}^{2}\big{]}
    +135a037[32a21(11934a213682a212b1217551a21b1226510b123)+440a21(337a21\displaystyle+135a_{03}^{7}\big{[}32a_{21}(11934a_{21}^{3}-682a_{21}^{2}b_{12}-17551a_{21}b_{12}^{2}-6510b_{12}^{3})+440a_{21}(337a_{21}
    363b12)(2a213b12)p122+21875(2a213b12)2p124]\displaystyle-363b_{12})(2a_{21}-3b_{12})p_{12}^{2}+21875(2a_{21}-3b_{12})^{2}p_{12}^{4}\big{]}
    +90a036a21(2a213b12)[32a21(14258a213+18094a212b1210465a21b1223906b123)\displaystyle+90a_{03}^{6}a_{21}(2a_{21}-3b_{12})\big{[}32a_{21}(14258a_{21}^{3}+18094a_{21}^{2}b_{12}-10465a_{21}b_{12}^{2}-3906b_{12}^{3})
    +8(71196a21340193a212b1294064a21b12232550b123)p122\displaystyle+8(71196a_{21}^{3}-40193a_{21}^{2}b_{12}-94064a_{21}b_{12}^{2}-32550b_{12}^{3})p_{12}^{2}
    +25(2584a215991b12)(2a213b12)p124]+12a035a212(2a213b12)2\displaystyle+25(2584a_{21}-5991b_{12})(2a_{21}-3b_{12})p_{12}^{4}\big{]}+12a_{03}^{5}a_{21}^{2}(2a_{21}-3b_{12})^{2}
    ×[48a21(42952a213+110952a212b12+67965a21b122+21700b123)\displaystyle\times\big{[}48a_{21}(42952a_{21}^{3}+110952a_{21}^{2}b_{12}+67965a_{21}b_{12}^{2}+21700b_{12}^{3})
    +80(38078a213+31157a212b1270341a21b12213020b123)p122+25(15556a212\displaystyle+80(38078a_{21}^{3}+31157a_{21}^{2}b_{12}-70341a_{21}b_{12}^{2}-13020b_{12}^{3})p_{12}^{2}+25(15556a_{21}^{2}
    77328a21b124509b122)p124]+8a034a213(2a213b12)3[144a21(5248a213\displaystyle-77328a_{21}b_{12}-4509b_{12}^{2})p_{12}^{4}\big{]}+8a_{03}^{4}a_{21}^{3}(2a_{21}-3b_{12})^{3}\big{[}144a_{21}(5248a_{21}^{3}
    +15748a212b12+13517a21b122+4340b123)+40(67916a213+164530a212b12\displaystyle+15748a_{21}^{2}b_{12}+13517a_{21}b_{12}^{2}+4340b_{12}^{3})+40(67916a_{21}^{3}+164530a_{21}^{2}b_{12}
    +39263a21b122+28644b123)p12215(19608a212+110966a21b12+242733b122)p124]\displaystyle+39263a_{21}b_{12}^{2}+28644b_{12}^{3})p_{12}^{2}-15(19608a_{21}^{2}+110966a_{21}b_{12}+242733b_{12}^{2})p_{12}^{4}\big{]}
    +16a033a214(2a213b12)4p122[8(53754a213+165299a212b12+113365a21b122+43400b123)\displaystyle+16a_{03}^{3}a_{21}^{4}(2a_{21}-3b_{12})^{4}p_{12}^{2}\big{[}8(53754a_{21}^{3}+165299a_{21}^{2}b_{12}\!+\!113365a_{21}b_{12}^{2}\!+\!43400b_{12}^{3})
    15(12172a212+25588a21b12+72391b122)p122]+32a032a215(2a213b12)5p122\displaystyle-15(12172a_{21}^{2}+25588a_{21}b_{12}+72391b_{12}^{2})p_{12}^{2}\big{]}+32a_{03}^{2}a_{21}^{5}(2a_{21}-3b_{12})^{5}p_{12}^{2}
    ×[8(2156a213+7415a212b12+6106a21b122+2170b123)5(6944a212+9626a21b12\displaystyle\times\big{[}8(2156a_{21}^{3}+7415a_{21}^{2}b_{12}+6106a_{21}b_{12}^{2}+2170b_{12}^{3})-5(6944a_{21}^{2}+9626a_{21}b_{12}
    +26097b122)p122]64a03a216(2a213b12)6p124(3276a212+3656a21b12+7205b122)\displaystyle+26097b_{12}^{2})p_{12}^{2}\big{]}-64a_{03}a_{21}^{6}(2a_{21}-3b_{12})^{6}p_{12}^{4}(3276a_{21}^{2}+3656a_{21}b_{12}+7205b_{12}^{2})
    128a217p124(2a213b12)7(104a212+102a21b12+145b122).\displaystyle-128a_{21}^{7}p_{12}^{4}(2a_{21}-3b_{12})^{7}(104a_{21}^{2}+102a_{21}b_{12}+145b_{12}^{2}).

    Since a03<12(b12a21)2a_{03}<-\frac{1}{2}(b_{12}-a_{21})^{2}, we have a032a21b12<0a_{03}-2a_{21}b_{12}<0. Then, to solve V7(ε)=0V_{7}(\varepsilon)=0, we only need to find the solutions to the equations: M4M5=V~7a=V~7b=V~7c=0M_{4}M_{5}=\widetilde{V}_{7a}=\widetilde{V}_{7b}=\widetilde{V}_{7c}=0. Thus, we compute the resultants of M4M_{4}, M5M_{5}, V~7a\widetilde{V}_{7a}, V~7b\widetilde{V}_{7b} and V~7c\widetilde{V}_{7c} with respect to a03a_{03}, respectively, and obtain

    Res[M4,V~7a,V~7b,V~7c,a03]\displaystyle\ {\rm Res}[M_{4},\widetilde{V}_{7a},\widetilde{V}_{7b},\widetilde{V}_{7c},a_{03}]
    =\displaystyle= a216(2a213b12)3(a21+b12)N4[C1,C2a215(2a213b12)3N3,\displaystyle\ a_{21}^{6}(2a_{21}-3b_{12})^{3}(a_{21}+b_{12})N_{4}\big{[}C_{1},\ C_{2}a_{21}^{5}(2a_{21}-3b_{12})^{3}N_{3},
    C3a219(2a213b12)6(a21+b12)N3],\displaystyle\hskip 153.21204ptC_{3}a_{21}^{9}(2a_{21}-3b_{12})^{6}(a_{21}+b_{12})N_{3}\big{]},
    Res[M5,V~7a,V~7b,V~7c,a03]\displaystyle\ {\rm Res}[M_{5},\widetilde{V}_{7a},\widetilde{V}_{7b},\widetilde{V}_{7c},a_{03}]
    =\displaystyle= a213(2a213b12)6(a21+b12)2N1[C4,C5a213(2a213b12)4N2,\displaystyle\ a_{21}^{3}(2a_{21}-3b_{12})^{6}(a_{21}+b_{12})^{2}N_{1}\big{[}C_{4},\ C_{5}a_{21}^{3}(2a_{21}-3b_{12})^{4}N_{2},
    C6a216(a211)(25+2a21)(2a213b12)8(a21+b12)2],\displaystyle\hskip 158.99377ptC_{6}a_{21}^{6}(a_{21}-1)(25+2a_{21})(2a_{21}-3b_{12})^{8}(a_{21}+b_{12})^{2}\big{]},

    where CiC_{i}, i=1,2,,6i=1,2,\cdots,6, are real constants, and

    N1=\displaystyle N_{1}= (2a213p122)(2a21+3p122),\displaystyle\ (2a_{21}-3p_{12}^{2})(2a_{21}+3p_{12}^{2}),
    N2=\displaystyle N_{2}= 2a21(2a21+3b12)2(8a21+3b12)(a21+6b12)p122,\displaystyle\ 2a_{21}(2a_{21}+3b_{12})^{2}-(8a_{21}+3b_{12})(a_{21}+6b_{12})p_{12}^{2},
    N3=\displaystyle N_{3}= 8a212(4a21+5b12)2+5a21(13a212+78a21b12+85b122)p122+50(a21+b12)2p124,\displaystyle\ 8a_{21}^{2}(4a_{21}+5b_{12})^{2}+5a_{21}(13a_{21}^{2}+78a_{21}b_{12}+85b_{12}^{2})p_{12}^{2}+50(a_{21}+b_{12})^{2}p_{12}^{4},
    N4=\displaystyle N_{4}= 12a212(4a21+5b12)2(107a21240a21b12175b122)\displaystyle\ 12a_{21}^{2}(4a_{21}+5b_{12})^{2}(107a_{21}^{2}-40a_{21}b_{12}-175b_{12}^{2})
    10a21(4682a214+4193a213b1216315a212b12223925a21b1237875b124)p122\displaystyle-10a_{21}(4682a_{21}^{4}+4193a_{21}^{3}b_{12}-16315a_{21}^{2}b_{12}^{2}-23925a_{21}b_{12}^{3}-7875b_{12}^{4})p_{12}^{2}
    +5(5954a2147331a213b1221660a212b122+33975a21b123+47250b124)p124.\displaystyle+5(5954a_{21}^{4}-7331a_{21}^{3}b_{12}-21660a_{21}^{2}b_{12}^{2}+33975a_{21}b_{12}^{3}+47250b_{12}^{4})p_{12}^{4}.

    To find the common factors of the above resultants, we only need to consider the following three subcases due to a21(2a213b12)(a21+b12)0a_{21}(2a_{21}-3b_{12})(a_{21}+b_{12})\neq 0.

    (ii-3-1-2-4-1) If a21=32p122a_{21}=\frac{3}{2}p_{12}^{2}, we obtain the simplified polynomials,

    M5=\displaystyle M_{5}= 9(a03b12p122+p124)(2a033b12p122+3p124),\displaystyle\ 9(a_{03}-b_{12}p_{12}^{2}+p_{12}^{4})(2a_{03}-3b_{12}p_{12}^{2}+3p_{12}^{4}),
    V~7a=\displaystyle\widetilde{V}_{7a}= 9(a03b12p122+p124)[70a03381p126(b12p122)2(5b123p122)\displaystyle\ 9(a_{03}-b_{12}p_{12}^{2}+p_{12}^{4})\big{[}70a_{03}^{3}-81p_{12}^{6}(b_{12}-p_{12}^{2})^{2}(5b_{12}-3p_{12}^{2})
    +54a03p124(b12p122)(10b1213p122)15a032p122(27b1219p122)].\displaystyle+54a_{03}p_{12}^{4}(b_{12}-p_{12}^{2})(10b_{12}-13p_{12}^{2})-15a_{03}^{2}p_{12}^{2}(27b_{12}-19p_{12}^{2})\big{]}.

    If a03=(b12p122)p122a_{03}=(b_{12}-p_{12}^{2})p_{12}^{2}, we have M5=V~7a=0M_{5}=\widetilde{V}_{7a}=0, which yields Δ+=14(2b12p122)2>0\Delta^{+}=\frac{1}{4}(2b_{12}-p_{12}^{2})^{2}>0 violating the condition Δ+<0\Delta^{+}<0. If a03=32(b12p122)p122a_{03}=\frac{3}{2}(b_{12}-p_{12}^{2})p_{12}^{2}, we obtain b12=32p122b_{12}=-\frac{3}{2}p_{12}^{2} from V~7a=0\widetilde{V}_{7a}=0, which contradicts the condition b12a21b_{12}\neq-a_{21}.

    (ii-3-1-2-4-2) Assume that a21=32p122a_{21}=-\frac{3}{2}p_{12}^{2}. Similar to the above process, we use the equation M5=V~7a=0M_{5}=\widetilde{V}_{7a}=0 with the constraint Δ+<0\Delta^{+}<0 to obtain one set of parameter values:

    {2a21+3p122=a033(b12+p122)p122=b21+2p12=p033p122=0}.\displaystyle\{2a_{21}+3p_{12}^{2}=a_{03}-3(b_{12}+p_{12}^{2})p_{12}^{2}=b_{21}+2p_{12}=p_{03}-3p_{12}^{2}=0\}.

    Eliminating the perturbed parameter p12p_{12} we obtain the condition IV.

    (ii-3-1-2-4-3) Now assume N4=0N_{4}=0 and p120p_{12}\neq 0. We solve the equations: M4=V~7a=V~7b=V~7c=0M_{4}=\widetilde{V}_{7a}=\widetilde{V}_{7b}=\widetilde{V}_{7c}=0 to obtain the solutions under which V8(ε)=0V_{8}(\varepsilon)=0 and the 99th generalized Lyapunov constant is simplified to

    V9(ε)=\displaystyle V_{9}(\varepsilon)= πp12ε13122305904640a037a213(2a213b12)3(a21+b12)3[256a032a212(2a213b12)2\displaystyle\ \frac{-\,\pi\,p_{12}\,\varepsilon^{13}}{122305904640a_{03}^{7}a_{21}^{3}(2a_{21}-3b_{12})^{3}(a_{21}+b_{12})^{3}}\big{[}256a_{03}^{2}a_{21}^{2}(2a_{21}-3b_{12})^{2}
    ×(a21+b12)2V9a+32a21(2a213b12)(a21+b12)V9bε23V9cε4],\displaystyle\times(a_{21}+b_{12})^{2}V_{9a}+32a_{21}(2a_{21}-3b_{12})(a_{21}+b_{12})V_{9b}\varepsilon^{2}-3V_{9c}\varepsilon^{4}\big{]},

    where V9aV_{9a}, V9bV_{9b} and V9cV_{9c} are polynomials in a03a_{03}, a21a_{21}, b12b_{12} and p12p_{12}, with 221221, 317317 and 299299 terms, respectively. We compute the resultants of M4M_{4}, V9aV_{9a}, V9bV_{9b} and M9cM_{9c} with respect to a03a_{03}, respectively, and obtain

    Res\displaystyle{\rm Res} [M4,V9a,V9b,V9c,a03]=[C7a2113(2a213b12)7(a21+b12)3N5,\displaystyle[M_{4},V_{9a},V_{9b},V_{9c},a_{03}]=\big{[}C_{7}a_{21}^{13}(2a_{21}-3b_{12})^{7}(a_{21}+b_{12})^{3}N_{5},
    C8a2119(2a213b12)11(a21+b12)3N6,C9a2121(2a213b12)13(a21+b12)3N3N7],\displaystyle\ C_{8}a_{21}^{19}(2a_{21}-3b_{12})^{11}(a_{21}+b_{12})^{3}N_{6},\ C_{9}a_{21}^{21}(2a_{21}-3b_{12})^{13}(a_{21}+b_{12})^{3}N_{3}N_{7}\big{]},

    where CiC_{i}, i=7,8,9i=7,8,9, are real constants, and N5N_{5}, N6N_{6} and N7N_{7} are polynomials in a21a_{21}, b12b_{12} and p12p_{12}, having 45 terms, 81 terms and 47 terms, respectively.

    We again calculate the resultants of N4N_{4}, N5N_{5}, N6N_{6}, N3N7N_{3}N_{7} with respect to a21a_{21}, respectively, to have

    Res[N4,N5,N6,N3N7,a21]=b1236(2b1225p122)2(10b12+11p122)p1232N8×[C10N9,C11(2b1225)(3150b12143)p124N10N11,C12(2b1225)(3150b12143)p124N10N12N13],\begin{array}[]{rl}{\rm Res}[N_{4},N_{5},N_{6},N_{3}N_{7},a_{21}]=\!\!\!\!&b_{12}^{36}(2b_{12}-25p_{12}^{2})^{2}(10b_{12}+11p_{12}^{2})p_{12}^{32}N_{8}\\ \!\!\!\!&\times\big{[}C_{10}N_{9},\ C_{11}(2b_{12}-25)(3150b_{12}-143)p_{12}^{4}N_{10}N_{11},\\ &\quad\,C_{12}(2b_{12}-25)(3150b_{12}-143)p_{12}^{4}N_{10}N_{12}N_{13}\big{]},\\ \end{array}

    where CiC_{i}, i=10,11,12i=10,11,12, are real constants, and

    N8=\displaystyle N_{8}= 86943853334p1212234718996369b12p1210+107204353618b122p128\displaystyle\,-86943853334p_{12}^{12}-234718996369b_{12}p_{12}^{10}+107204353618b_{12}^{2}p_{12}^{8}
    +414472065816b123p12688238153808b124p1242186887248b125p122+792148896b126,\displaystyle+414472065816b_{12}^{3}p_{12}^{6}-88238153808b_{12}^{4}p_{12}^{4}-2186887248b_{12}^{5}p_{12}^{2}+792148896b_{12}^{6},

    while the polynomials Ni(b12)N_{i}(b_{12}), i=9,12,,13i=9,12,\cdots,13, do not have common nonzero roots. So, we only have possible solutions from (2b1225p122)2(10b12+11p122)N8(2b_{12}-25p_{12}^{2})^{2}(10b_{12}+11p_{12}^{2})N_{8}. If b12=252p122b_{12}=\frac{25}{2}p_{12}^{2}, we solve N4=N5=0N_{4}=N_{5}=0 to obtain a21=252p122a_{21}=-\frac{25}{2}p_{12}^{2}, which yields a21+b12=0a_{21}+b_{12}=0, contradicting the condition a21+b120a_{21}+b_{12}\neq 0. If b12=1110p122b_{12}=-\frac{11}{10}p_{12}^{2}, we solve M4=N4=N5=0M_{4}=N_{4}=N_{5}=0 to have a21=32p122a_{21}=\frac{3}{2}p_{12}^{2}, and either a03=910p122a_{03}=-\frac{9}{10}p_{12}^{2} or a03=2110p122a_{03}=-\frac{21}{10}p_{12}^{2}, which violates Δ+<0\Delta^{+}<0.

    Now, assume that N8=0N_{8}=0. We have V9(ε)=V10(ε)=0V_{9}(\varepsilon)=V_{10}(\varepsilon)=0, and obtain the ε17\varepsilon^{17}-order term V11aV_{11a} in the 1111th generalized Lyapunov constant, which is a polynomial in a03a_{03}, a21a_{21} and b12b_{12} having 237237 terms. We compute the resultant of M4M_{4} and V11aV_{11a} with respect to a03a_{03} to obtain

    Res[M4,V11a,a03]=C13a2119(2a213b12)11(a21+b12)4N14,\displaystyle{\rm Res}[M_{4},V_{11a},a_{03}]=C_{13}a_{21}^{19}(2a_{21}-3b_{12})^{11}(a_{21}+b_{12})^{4}N_{14},

    where C13C_{13} is a real constant, and N14N_{14} is a polynomial in a21a_{21} and b12b_{12} having 179179 terms. We again calculate the resultant of N4N_{4} and N14N_{14} with respect to a21a_{21}, and have

    Res[N4,N14,a21]\displaystyle{\rm Res}[N_{4},N_{14},a_{21}] =C14b1248p1236(2b1225p122)3(10b12+11p122)N15,\displaystyle=C_{14}b_{12}^{48}p_{12}^{36}(2b_{12}-25p_{12}^{2})^{3}(10b_{12}+11p_{12}^{2})N_{15},

    where C14C_{14} is a real constant, and N15N_{15} is a 7676th-degree polynomial in b12b_{12} and p12p_{12}. Since the resultant of N8N_{8} and N15N_{15} with respect to b12b_{12} equals C15p124560C_{15}p_{12}^{456}\neq 0, where C15C_{15} is a real constant, we conclude that there do not exist parameter values such that all εk\varepsilon^{k}-order terms in V9(ε)V_{9}(\varepsilon), V10(ε)V_{10}(\varepsilon) and V11(ε)V_{11}(\varepsilon) vanish.

    (ii-3-2) Since q02=0q_{02}=0 we assume that b21=13p12p21(2p211)b_{21}=\frac{1}{3}p_{12}p_{21}(2p_{21}-1) and 1+p2101+p_{21}\neq 0 based on V5d=0V_{5d}=0. If p12=0p_{12}=0, we have V5(ε)=0V_{5}(\varepsilon)=0, which is included in the condition II. Hence, we assume that p120p_{12}\neq 0. Similar to the analysis for the case (ii-3-1), we consider the two subcases b12=a21b_{12}=-a_{21} and M3=0M_{3}=0.

    (ii-3-2-1) Under the condition b12=a21b_{12}=-a_{21}, we assume that V5b=V5c=0V_{5b}=V_{5c}=0, which yields V5(ε)=V6(ε)=0V_{5}(\varepsilon)=V_{6}(\varepsilon)=0. Then, setting the ε17\varepsilon^{17}-order term in V7(ε)V_{7}(\varepsilon) zero we obtain 516p21+4p212=05-16p_{21}+4p_{21}^{2}=0, under which the polynomials V5bV_{5b} and V5cV_{5c} are simplified to

    V^5b=\displaystyle\widehat{V}_{5b}= 27a03324a212+940a21p1223(153a03+54a212+916a21p122)p21,\displaystyle-27a_{03}-324a_{21}^{2}+940a_{21}p_{12}^{2}-3(153a_{03}+54a_{21}^{2}+916a_{21}p_{12}^{2})p_{21},
    V^5c=\displaystyle\widehat{V}_{5c}= 243a21+54p03+2495p122+6(378a21+153p031249p122)p21.\displaystyle\ 243a_{21}+54p_{03}+2495p_{12}^{2}+6(378a_{21}+153p_{03}-1249p_{12}^{2})p_{21}.

    It is easy to find from V^5b=V^5c=0\widehat{V}_{5b}=\widehat{V}_{5c}=0 that

    a03=\displaystyle a_{03}= 2a21[162a21470p122+3(27a21+458p122)p21]27(1+17p21)<0,(due to (27))\displaystyle-\frac{2a_{21}[162a_{21}-470p_{12}^{2}+3(27a_{21}+458p_{12}^{2})p_{21}]}{27(1+17p_{21})}<0,\ \ (\textrm{due to \eqref{Eqn27}})
    p03=\displaystyle p_{03}= 243a21+2495p122+6(378a211249p122)p2154(1+17p21).\displaystyle-\frac{243a_{21}+2495p_{12}^{2}+6(378a_{21}-1249p_{12}^{2})p_{21}}{54(1+17p_{21})}.

    Further, we have the following simplified polynomials:

    V^7a=\displaystyle\widehat{V}_{7a}= 1458(2034p21695)a21215(153111a2125780708p122)p122\displaystyle\ 1458(2034p_{21}-695)a_{21}^{2}-15(153111a_{21}-25780708p_{12}^{2})p_{12}^{2}
    +2(3360789a21565883302p122)p122p21,\displaystyle+2(3360789a_{21}-565883302p_{12}^{2})p_{12}^{2}p_{21},
    V^7b=\displaystyle\widehat{V}_{7b}= 2916(16169p215525)a212+5(598986a21278284327p122)p122\displaystyle\ 2916(16169p_{21}-5525)a_{21}^{2}+5(598986a_{21}-278284327p_{12}^{2})p_{12}^{2}
    18(487243a21226233559p122)p122p21.\displaystyle-18(487243a_{21}-226233559p_{12}^{2})p_{12}^{2}p_{21}.

    Then, we compute the resultant of V^7a\widehat{V}_{7a} and V^7b\widehat{V}_{7b} with respect to p21p_{21} to obtain

    Res\displaystyle{\rm Res} [V^7a,V^7b,p12]= 871696100250000a218(1979853717341274735024025\displaystyle[\widehat{V}_{7a},\widehat{V}_{7b},p_{12}]=\ 871696100250000a_{21}^{8}(1979853717341274735024025
    23176897174962961258321510p21+101743987046015667217950816p212\displaystyle-23176897174962961258321510p_{21}+101743987046015667217950816p_{21}^{2}
    198508767310489076670643400p213+145238548628125887939760784p214)20,\displaystyle-198508767310489076670643400p_{21}^{3}+145238548628125887939760784p_{21}^{4})^{2}\neq 0,

    under the conditions: 516p21+4p212=05-16p_{21}+4p_{21}^{2}=0 and a210a_{21}\neq 0. This shows that there are no parameter values satisfying V^7a=V^7b=0\widehat{V}_{7a}=\widehat{V}_{7b}=0.

    (ii-3-2-2) Assume b12a21b_{12}\neq-a_{21}. Letting the ε17\varepsilon^{17}-order term in V7(ε)V_{7}(\varepsilon) be zero we have 516p21+4p212=05-16p_{21}+4p_{21}^{2}=0, yielding the following simplified polynomials:

    V¯5a=\displaystyle\overline{V}_{5a}= (5+28p21)a03+6(2a213b12)a21,\displaystyle\ (5+28p_{21})a_{03}+6(2a_{21}-3b_{12})a_{21},
    V¯5c=\displaystyle\overline{V}_{5c}= 1944a21+3726b122295p03+18735p122\displaystyle-1944a_{21}+3726b_{12}-2295p_{03}+18735p_{12}^{2}
    +(7452p03+6156a2112474b1254962p122)p21.\displaystyle+(7452p_{03}+6156a_{21}-12474b_{12}-54962p_{12}^{2})p_{21}.

    Setting V¯5a=V¯5c=0\overline{V}_{5a}=\overline{V}_{5c}=0 gives

    a03=\displaystyle a_{03}= 6a21(2a213b12)5+28p21,\displaystyle-\frac{6a_{21}(2a_{21}-3b_{12})}{5+28p_{21}},
    p03=\displaystyle p_{03}= 1944a213726b1218735p122+(12474b126156a21)p21+54962p12227(276p2185).\displaystyle\frac{1944a_{21}-3726b_{12}-18735p_{12}^{2}+(12474b_{12}-6156a_{21})p_{21}+54962p_{12}^{2}}{27(276p_{21}-85)}.

    We use the above solutions to simplify V¯5b\overline{V}_{5b} and V7(ε)V_{7}(\varepsilon) to obtain the following polynomials:

    V¯5b=\displaystyle\overline{V}_{5b}= 324a212(6222p212185)+2a21[p122(86250505252436996p21)+810b12(12636p21\displaystyle\ 324a_{21}^{2}(6222p_{21}\!-\!2185)\!+\!2a_{21}\big{[}p_{12}^{2}(86250505\!-\!252436996p_{21})\!+\!810b_{12}(12636p_{21}
    4345)]80b12[p122(637379518653564p21)+810b12(1204p21411)],\displaystyle-\!4345)\big{]}-80b_{12}\big{[}p_{12}^{2}(6373795-18653564p_{21})+810b_{12}(1204p_{21}-411)\big{]},
    V¯7a=\displaystyle\overline{V}_{7a}= 1296a214(52553294p2118022295)27920b123[p122(6847437152004000728p21)\displaystyle 1296a_{21}^{4}(52553294p_{21}-18022295)-27920b_{12}^{3}\big{[}p_{12}^{2}(684743715-2004000728p_{21})
    +4050b12(25872p218839)]+4a213[p122(15991130051454680077698534p21)\displaystyle+4050b_{12}(25872p_{21}-8839)\big{]}+4a_{21}^{3}\big{[}p_{12}^{2}(1599113005145-4680077698534p_{21})
    +81b12(2484285362p21850029185)]+4a212b12[81b12(12255595953592556264p21)\displaystyle+81b_{12}(2484285362p_{21}\!-\!850029185)\big{]}+4a_{21}^{2}b_{12}\big{[}81b_{12}(1225559595\!-\!3592556264p_{21})
    +p122(5277460809917p211803276434885)]+2a21b122[810b12(918918755\displaystyle+p_{12}^{2}(5277460809917p_{21}-1803276434885)\big{]}+2a_{21}b_{12}^{2}\big{[}810b_{12}(918918755
    2690204032p21)p122(1470804241743543045101045752p21)],\displaystyle-2690204032p_{21})-p_{12}^{2}(14708042417435-43045101045752p_{21})\big{]},
    V¯7b=\displaystyle\overline{V}_{7b}= 729a213(24810684334p218478232095)80b12[9b12p122(8334257322035\displaystyle\ 729a_{21}^{3}(24810684334p_{21}-8478232095)-80b_{12}\big{[}9b_{12}p_{12}^{2}(8334257322035
    24391452032792p21)+10p124(11073747103453240898045749p21)\displaystyle-24391452032792p_{21})+10p_{12}^{4}(1107374710345-3240898045749p_{21})
    +3645b122(3097706668p211058445047)]+9a212[p122(240122211943205\displaystyle+3645b_{12}^{2}(3097706668p_{21}-1058445047)\big{]}+9a_{21}^{2}\big{[}p_{12}^{2}(240122211943205
    702753773748111p21)+486b12(61045680279p2120858985470)]\displaystyle-702753773748111p_{21})+486b_{12}(61045680279p_{21}-20858985470)\big{]}
    +a21[5p124(111885081254445327448472749544p21)+7290b122(30073269035\displaystyle+a_{21}\big{[}5p_{12}^{4}(111885081254445-327448472749544p_{21})+7290b_{12}^{2}(30073269035
    88014820004p21)9b12p122(5036433406353951473987501858484p21)],\displaystyle-88014820004p_{21})-9b_{12}p_{12}^{2}(503643340635395-1473987501858484p_{21})\big{]},
    V¯7c=\displaystyle\overline{V}_{7c}= 9234a212(333894386p21114095455)+3a21[p122(414050132594835\displaystyle\ 9234a_{21}^{2}(333894386p_{21}-114095455)+3a_{21}\big{[}p_{12}^{2}(414050132594835
    1211779848284032p21)810b12(2350803947568799585032p21)]\displaystyle-1211779848284032p_{21})-810b_{12}(23508039475-68799585032p_{21})\big{]}
    +40[350p124(1537000162544982613958p21)+2430b122(1831770089\displaystyle+40\big{[}350p_{12}^{4}(15370001625-44982613958p_{21})+2430b_{12}^{2}(1831770089
    5360953016p21)3b12p122(3186299869146593251840991288p21)],\displaystyle-5360953016p_{21})-3b_{12}p_{12}^{2}(31862998691465-93251840991288p_{21})\big{]},
    V¯7d=\displaystyle\overline{V}_{7d}= 10p122(43443127132p21)+648b12(261764p21)81a21(7072068p21).\displaystyle 10p_{12}^{2}(43443-127132p_{21})+648b_{12}(261-764p_{21})-81a_{21}(707-2068p_{21}).

    By computing the Groebner basis for 516p21+4p2125-16p_{21}+4p_{21}^{2}, V¯5b\overline{V}_{5b}, V¯7a\overline{V}_{7a}, V¯7b\overline{V}_{7b}, V¯7c\overline{V}_{7c} and V¯7d\overline{V}_{7d}, we obtain two polynomial equations:

    p122(1812510b12+13187705p12211105856p122p21)\displaystyle p_{12}^{2}(1812510b_{12}+13187705p_{12}^{2}-11105856p_{12}^{2}p_{21}) =0,\displaystyle=0,
    8667a21+12312b124030p12210368b12p2116880p122p21\displaystyle 8667a_{21}+12312b_{12}-4030p_{12}^{2}-10368b_{12}p_{21}-16880p_{12}^{2}p_{21} =0,\displaystyle=0,

    which yields

    a21=2290907855p122(1066162176p2122248798664p21+1571031845),\displaystyle a_{21}=\frac{2}{290907855}\,p_{12}^{2}(1066162176p_{21}^{2}-2248798664p_{21}+1571031845),
    b12=11812510p122(11105856p2113187705),\displaystyle b_{12}=\frac{1}{1812510}\,p_{12}^{2}(11105856p_{21}-13187705),

    under which V8(ε)=0V_{8}(\varepsilon)=0, and the 99th generalized Lyapunov constant is simplified as

    V9(ε)=\displaystyle V_{9}(\varepsilon)= πp12ε13C15(5+28p21)3(276p2185)3\displaystyle\ \frac{\pi p_{12}\varepsilon^{13}}{C_{15}(5+28p_{21})^{3}(276p_{21}-85)^{3}}
    ×[8p1210(479401837151164960733153380166419352018673132763343471p21\displaystyle\times\big{[}8p_{12}^{10}(479401837151164960733153380166419352018673132763343471p_{21}
    163805665483591789907309292300795181773767007636854130)\displaystyle\qquad\qquad-163805665483591789907309292300795181773767007636854130)
    +C16p128(35989441277122180465488301421620623564108062636p21\displaystyle\quad\ +C_{16}\,p_{12}^{8}(35989441277122180465488301421620623564108062636p_{21}
    12297145988872641623638918735129812334369100955)ε2\displaystyle\qquad\qquad-12297145988872641623638918735129812334369100955)\varepsilon^{2}
    +C17p126(28047246207962034726462002167497642752864p21\displaystyle\quad\ +C_{17}p_{12}^{6}(28047246207962034726462002167497642752864p_{21}
    9583396378659769296319246478020611770295)ε4\displaystyle\qquad\qquad-9583396378659769296319246478020611770295)\varepsilon^{4}
    +C18p124(317522879623822832097089379861158p21\displaystyle\quad\ +C_{18}p_{12}^{4}(317522879623822832097089379861158p_{21}
    108493632214946857667054466000365)ε6\displaystyle\qquad\qquad-108493632214946857667054466000365)\varepsilon^{6}
    +C19p122(1459616804250815596262396p21498732969803156525142505)ε8\displaystyle\quad\ +C_{19}p_{12}^{2}(1459616804250815596262396p_{21}-498732969803156525142505)\varepsilon^{8}
    +(2716325631250436p21928134798864155)ε10],\displaystyle\quad\ +(2716325631250436p_{21}-928134798864155)\varepsilon^{10}\big{]},

    where CiC_{i}, i=15,16,,19i=15,16,\cdots,19, are integers. Hence, any εk\varepsilon^{k}-order term in V9(ε)V_{9}(\varepsilon) is non-zero under the condition 516p21+4p212=05-16p_{21}+4p_{21}^{2}=0 and p120p_{12}\neq 0.

We have shown that the four conditions I, II, III and IV in Theorem 2.2 are necessary for the singular points (±1,0)(\pm 1,0) of system (9) to be nilpotent bi-center. Now, we prove that these four conditions are also sufficient.

If the condition I in Theorem 2.2 holds, system (9) is reduced to

(x˙y˙)={(2a21xy+a21x2y3b03xy2+a03y3,x+32x2a21y2+12x3a21xy2+b03y3,),ify>0,(2a21xy+a21x2y6b03y23b03xy2+a03y3,x+32x2a21y2+12x3a21xy2+b03y3,),ify<0,\left(\begin{array}[]{cc}\dot{x}\\ \dot{y}\end{array}\right)=\left\{\begin{aligned} &\left(\begin{aligned} &2a_{21}xy+a_{21}x^{2}y-3b_{03}xy^{2}+a_{03}y^{3},\\ &x+\tfrac{3}{2}x^{2}-a_{21}y^{2}+\tfrac{1}{2}x^{3}-a_{21}xy^{2}+b_{03}y^{3},\end{aligned}\right),&\text{if}\ \ y>0,\\[4.30554pt] &\left(\begin{aligned} &2a_{21}xy+a_{21}x^{2}y-6b_{03}y^{2}-3b_{03}xy^{2}+a_{03}y^{3},\\ &x+\tfrac{3}{2}x^{2}-a_{21}y^{2}+\tfrac{1}{2}x^{3}-a_{21}xy^{2}+b_{03}y^{3},\end{aligned}\right),&\text{if}\ \ y<0,\end{aligned}\right. (31)

via xx+1x\rightarrow x+1. The upper and the lower systems in (31) are Hamiltonian systems, having respectively the Hamiltonian functions,

H+(x,y)\displaystyle H^{+}(x,y) =12x212x318x4+a212x2y2b03xy3+a21xy2+a034y4,\displaystyle=-\frac{1}{2}x^{2}-\frac{1}{2}x^{3}-\frac{1}{8}x^{4}+\frac{a_{21}}{2}x^{2}y^{2}-b_{03}xy^{3}+a_{21}xy^{2}+\frac{a_{03}}{4}y^{4}, (32)
H(x,y)\displaystyle H^{-}(x,y) =12x212x318x4+a212x2y2b03xy3+a21xy22b03y3+a034y4,\displaystyle=-\frac{1}{2}x^{2}-\frac{1}{2}x^{3}-\frac{1}{8}x^{4}+\frac{a_{21}}{2}x^{2}y^{2}-b_{03}xy^{3}+a_{21}xy^{2}-2b_{03}y^{3}+\frac{a_{03}}{4}y^{4},

which shows that the condition, H+(x,0)H(x,0)H^{+}(x,0)\equiv H^{-}(x,0), in Proposition 2.1 of [10] is satisfied, and so the origin of system (31) is a center. Hence, system (9) has a nilpotent bi-center at (±1\pm 1,0), which consists of a monodromic singular point and a cusp.

If the condition II in Theorem 2.2 holds, the system (9) is simplified into a smooth one,

(x˙y˙)=(a21y+a21x2y+a03y3,x2+x32+b12xy2,),\left(\begin{array}[]{cc}\dot{x}\\ \dot{y}\end{array}\right)=\left(\begin{aligned} &-a_{21}y+a_{21}x^{2}y+a_{03}y^{3},\\[0.0pt] &-\tfrac{x}{2}+\tfrac{x^{3}}{2}+b_{12}xy^{2},\end{aligned}\right), (33)

which is symmetric with respect to the xx-axis, so we know that the singular points (±1\pm 1,0) of system (9) are bi-center.

If the condition III in Theorem 2.2 holds, system (9) again becomes a smooth one:

(x˙y˙)=(a21y+a21x2y+a03y3,x2+x32b21y+b21x2ya21xy223a21b21y3,),\left(\begin{array}[]{cc}\dot{x}\\ \dot{y}\end{array}\right)=\left(\begin{aligned} &-a_{21}y+a_{21}x^{2}y+a_{03}y^{3},\\ &-\tfrac{x}{2}+\tfrac{x^{3}}{2}-b_{21}y+b_{21}x^{2}y-a_{21}xy^{2}-\tfrac{2}{3}a_{21}b_{21}y^{3},\end{aligned}\right), (34)

which has the algebraic integral curve,

I(x,y)=\displaystyle I(x,y)= 3(9a03+2a212)(x2+2b21xy2a21y2)\displaystyle 3(9a_{03}+2a_{21}^{2})(x^{2}+2b_{21}xy-2a_{21}y^{2})
(3a032a212)(3x4+6b21x3y12a21x2y24a21b21xy36a03y4),\displaystyle-(3a_{03}-2a_{21}^{2})(3x^{4}+6b_{21}x^{3}y-12a_{21}x^{2}y^{2}-4a_{21}b_{21}xy^{3}-6a_{03}y^{4}),

giving an inverse integrating factor I(x,y)I(x,y) for this system. Thus, system (9) has a nilpotent bi-center at (±1\pm 1,0).

If the condition IV in Theorem 2.2 holds, system (9) is a smooth one:

(x˙y˙)=(38b212y38b212x2y+a03y3,x2b21y+x32+b21x2y+16a033b21412b212xy232a039b21424b21y3,).\left(\begin{array}[]{cc}\dot{x}\\ \dot{y}\end{array}\right)=\left(\begin{aligned} &\tfrac{3}{8}b_{21}^{2}y-\tfrac{3}{8}b_{21}^{2}x^{2}y+a_{03}y^{3},\\ &-\tfrac{x}{2}-b_{21}y+\tfrac{x^{3}}{2}+b_{21}x^{2}y+\tfrac{16a_{03}-3b_{21}^{4}}{12b_{21}^{2}}xy^{2}-\tfrac{32a_{03}-9b_{21}^{4}}{24b_{21}}y^{3},\end{aligned}\right). (35)

Actually, for an arbitrary nonzero constant rr, by the transformation,

x=X,y=rY,t=rT,x=X,\qquad y=rY,\qquad t=rT,

system (35) can be changed to

(X˙Y˙)=(38(b21r)2Y38(b21r)2X2Y+a03r4Y3X2b21rY+X32+b21rX2Y+16a03r43(b21r)412(b21r)2XY232a03r49(b21r)424(b21r)Y3),\left(\!\begin{array}[]{cc}\dot{X}\\ \dot{Y}\end{array}\!\right)\!=\!\left(\begin{aligned} &\tfrac{3}{8}(b_{21}r)^{2}Y-\tfrac{3}{8}(b_{21}r)^{2}X^{2}Y+a_{03}r^{4}Y^{3}\\[2.15277pt] &-\tfrac{X}{2}-b_{21}rY+\tfrac{X^{3}}{2}+b_{21}rX^{2}Y+\tfrac{16a_{03}r^{4}-3(b_{21}r)^{4}}{12(b_{21}r)^{2}}XY^{2}-\tfrac{32a_{03}r^{4}-9(b_{21}r)^{4}}{24(b_{21}r)}Y^{3}\end{aligned}\right)\!, (36)

which implies that system (35) is invariant under the following transformation (a03,b21)(a03r4,b21r)(a_{03},b_{21})\rightarrow(a_{03}r^{4},b_{21}r). Hence, we can always choose proper rr to satisfy 𝔣3+=2a03316b214=2\mathfrak{f}_{3}^{+}=2a_{03}-\frac{3}{16}b_{21}^{4}=-2, yielding a03=332b2121a_{03}=\frac{3}{32}b_{21}^{2}-1. Then, the origin of system (35) is a center according to the condition C3C_{3} of Theorem 4.2 in [32].

Refer to caption
Refer to caption
Refer to caption
Refer to caption
Figure 3: The phase portraits of system (9) showing bi-center at (±1,0)(\pm 1,0) for (a) Condition I: a21=1,a12=3,a03=3,b02=b21=0,a02=b12=b03=1a_{21}=1,\,a_{12}=3,\,a_{03}=-3,\,b_{02}=b_{21}=0,\,a_{02}=b_{12}=b_{03}=-1; (b) Condition II: a21=a03=1,a02=a12=b02=b21=b12=b03=0a_{21}=a_{03}=-1,\,a_{02}=a_{12}=b_{02}=b_{21}=b_{12}=b_{03}=0; (c) Condition III: a21=1,a03=4,a02=a12=b02=0,b21=3510,b12=1,b03=55a_{21}=1,\,a_{03}=-4,\,a_{02}=a_{12}=b_{02}=0,\,b_{21}=\frac{3\sqrt{5}}{10},\,b_{12}=-1,\,b_{03}=-\frac{\sqrt{5}}{5}; (d) Condition IV: a21=32,a03=3,a02=a12=b02=0,b21=2,b12=2,b03=5a_{21}=-\frac{3}{2},\,a_{03}=-3,\,a_{02}=a_{12}=b_{02}=0,\,b_{21}=-2,\,b_{12}=-2,\,b_{03}=5.
Example 4.2.

The global phase portraits of system (9) corresponding to the four bi-center conditions I, II, III and IV, with respectively three sets of parameter values, show the bi-center at (±1\pm 1,0), as illustrated in Figure 3.

4.2 The 22nd-order critical point (1,0)(1,0) of the upper system in (9)

In this subsection, we consider the center conditions associated with (1,0)(1,0) of the first system of (9) with multiplicity two. Thus, assume that 𝔣2+<0\mathfrak{f}_{2}^{+}<0, i.e., a02+a12<0a_{02}+a_{12}<0. Then, the singular point (1,0)(1,0) in the upper smooth system of (9) is a cusp. If 𝔣2=a12a02=0\mathfrak{f}_{2}^{-}=a_{12}-a_{02}=0, then (1,0)(1,0) in the lower system of (9) is a 3rd-order critical point. If 𝔣2=a12a020\mathfrak{f}_{2}^{-}=a_{12}-a_{02}\neq 0, (1,0)(1,0) in the lower smooth system of (9) is also a cusp. Similarly, recall that the singular point (1,01,0) cannot be a monodromic singular point when a02a12>0a_{02}-a_{12}>0, and hence we only consider the case when a02a120a_{02}-a_{12}\leq 0. Therefore, combining a02+a12<0a_{02}+a_{12}<0 and a02a120a_{02}-a_{12}\leq 0 we have either a02=a12<0a_{02}=a_{12}<0 or a02+|a12|<0a_{02}+|a_{12}|<0.

Example 4.3.

The global phase portrait of system (9) with a02=2a_{02}=2, a12=3a_{12}=-3 and a03=a21=b02=b12=b21=b03=1a_{03}=a_{21}=b_{02}=b_{12}=b_{21}=b_{03}=1 shows that the nilpotent singular points (±1,0)(\pm 1,0) are two cusps, see Figure 4.

Refer to caption
Figure 4: The phase portrait of system (9) with a02=2a_{02}=2, a12=3a_{12}=-3 and a03=a21=b02=b12=b21=b03=1a_{03}=a_{21}=b_{02}=b_{12}=b_{21}=b_{03}=1, showing two cusps at (±1,0)(\pm 1,0).

We apply the generalized Poincaré-Lyapunov method again to system (9) with either a02=a12<0a_{02}=a_{12}<0 or a02+|a12|<0a_{02}+|a_{12}|<0. Introducing the transformation (x,y,t)(ε3(x+1),ε2y,tε)(x,y,t)\rightarrow(\varepsilon^{3}(x+1),\varepsilon^{2}y,\frac{t}{\varepsilon}) into system (9), we obtain the perturbed system,

(x˙y˙)={(y+2(a21+p21ε2)εxy+(a02+a12+p02ε2)y2+(a21+p21ε2)ε4x2y+(a12+p12ε2)ε3xy2+(a03+p03ε2)ε2y3,x+32ε3x2+12ε6x3+(b02+b12+q02ε2)εy2+2(b21+q21ε2)ε2xy+(b21+ε2)ε5x2y+(b12+q12ε2)ε4xy2+(b03+q03ε2)ε3y3,),ify>0,(y+2(a21+p21ε2)εxy+(a21+p21ε2)ε4x2y(a02a12+p02ε22p12ε2)y2+(a12+p12ε2)ε3xy2+(a03+p03ε2)ε2y3,x+32εx2+12ε6x3(b02b12+q02ε22q12ε2)εy2+2(b21+q21ε2)ε2xy+(b21+q21ε2)ε5x2y+(b12+q12ε2)ε4xy2+(b03+q03ε2)ε3y3,),ify<0.\left(\begin{array}[]{cc}\dot{x}\\ \dot{y}\end{array}\right)=\left\{\begin{aligned} &\left(\begin{aligned} &-y+2(a_{21}+p_{21}\varepsilon^{2})\varepsilon xy+(a_{02}+a_{12}+p_{02}\varepsilon^{2})y^{2}\\ &+(a_{21}+p_{21}\varepsilon^{2})\varepsilon^{4}x^{2}y+(a_{12}+p_{12}\varepsilon^{2})\varepsilon^{3}xy^{2}\\ &+(a_{03}+p_{03}\varepsilon^{2})\varepsilon^{2}y^{3},\\[4.30554pt] &x+\tfrac{3}{2}\varepsilon^{3}x^{2}+\tfrac{1}{2}\varepsilon^{6}x^{3}+(b_{02}+b_{12}+q_{02}\varepsilon^{2})\varepsilon y^{2}\\ &+2(b_{21}+q_{21}\varepsilon^{2})\varepsilon^{2}xy+(b_{21}+\varepsilon^{2})\varepsilon^{5}x^{2}y\\ &+(b_{12}+q_{12}\varepsilon^{2})\varepsilon^{4}xy^{2}+(b_{03}+q_{03}\varepsilon^{2})\varepsilon^{3}y^{3},\end{aligned}\right),&\text{if}\ \ y>0,\\[4.30554pt] &\left(\begin{aligned} &-y+2(a_{21}+p_{21}\varepsilon^{2})\varepsilon xy+(a_{21}+p_{21}\varepsilon^{2})\varepsilon^{4}x^{2}y\\ &-(a_{02}-a_{12}+p_{02}\varepsilon^{2}-2p_{12}\varepsilon^{2})y^{2}\\ &+(a_{12}+p_{12}\varepsilon^{2})\varepsilon^{3}xy^{2}+(a_{03}+p_{03}\varepsilon^{2})\varepsilon^{2}y^{3},\\[4.30554pt] &x+\tfrac{3}{2}\varepsilon x^{2}+\tfrac{1}{2}\varepsilon^{6}x^{3}-(b_{02}-b_{12}+q_{02}\varepsilon^{2}-2q_{12}\varepsilon^{2})\varepsilon y^{2}\\ &+2(b_{21}+q_{21}\varepsilon^{2})\varepsilon^{2}xy+(b_{21}+q_{21}\varepsilon^{2})\varepsilon^{5}x^{2}y\\ &+(b_{12}+q_{12}\varepsilon^{2})\varepsilon^{4}xy^{2}+(b_{03}+q_{03}\varepsilon^{2})\varepsilon^{3}y^{3},\end{aligned}\right),&\text{if}\ \ y<0.\end{aligned}\right. (37)

which satisfies the condition (29) for the perturbed parameters.

Similarly, we first present a table to outline the proof for the center conditions V and VI.

Table 2: Outline of the proof for the center conditions V-VI.
Cases Conditions
(i) a02+|a02|<0a_{02}+|a_{02}|<0 (i-1) a12=0a_{12}=0 (i-1-1) q21=0q_{21}=0 (i-1-1-1) p12=b212p_{12}=\tfrac{b_{21}}{2} V
(i-1-1-2) p12b212p_{12}\neq\tfrac{b_{21}}{2} V
(i-1-2) q12=1q_{12}=-1 (i-1-2-1) b12=a21b_{12}=-a_{21} V
(i-1-2-2) p12=b212p_{12}=\tfrac{b_{21}}{2}
(i-2) a120a_{12}\neq 0 (i-2-1) q21=0q_{21}=0 VI
(i-2-2) q12=1q_{12}=-1 VI
(ii) a02=a12<0a_{02}=a_{12}<0 VI

The first two generalized Lyapunov constants at the origin of system (37) are

V1(ε)=0andV2(ε)=83ε[b02+(q02q12)ε2].V_{1}(\varepsilon)=0\quad\textrm{and}\quad V_{2}(\varepsilon)=\frac{8}{3}\varepsilon\big{[}b_{02}+(q_{02}-q_{12})\varepsilon^{2}\big{]}.

Setting V2(ε)=0V_{2}(\varepsilon)=0 we get the necessary center condition: b02=0b_{02}=0, q02=q12q_{02}=q_{12}, for system (37). Then, it follows from the discussion given at the beginning of this subsection that we consider two cases: (i) a02+|a12|<0a_{02}+|a_{12}|<0 and (ii) a02=a12<0a_{02}=a_{12}<0.

  1. (i)

    By the ε\varepsilon-order term of V3(ε)V_{3}(\varepsilon), we have two subcases (i-1) a12=0a_{12}=0 and (i-2) a120a_{12}\neq 0.

    (i-1) If a12=0a_{12}=0, we have a02<0a_{02}<0. Then, the 33rd generalized Lyapunov constant is given by

    V3(ε)=\displaystyle V_{3}(\varepsilon)= π4ε3{3b032b12b21+2(a21+b12)p122(1+q12)q21ε4\displaystyle-\frac{\pi}{4}\varepsilon^{3}\big{\{}3b_{03}-2b_{12}b_{21}+2(a_{21}+b_{12})p_{12}-2(1+q_{12})q_{21}\varepsilon^{4}
    +[3q032b21(1+q12)+(1+2p21+2p12)p122b12q21]ε2}.\displaystyle\qquad\quad+\big{[}3q_{03}-2b_{21}(1+q_{12})+(1+2p_{21}+2p_{12})p_{12}-2b_{12}q_{21}\big{]}\varepsilon^{2}\big{\}}.

    By setting the ε7\varepsilon^{7}-order term in V3(ε)V_{3}(\varepsilon) zero we have the following two subcases: (i-1-1) q21=0q_{21}=0 and (i-1-2) q12=1q_{12}=-1.

    (i-1-1) Assume that q21=0q_{21}=0. Letting the ε3\varepsilon^{3}-order and ε5\varepsilon^{5}-order terms in V3(ε)=0V_{3}(\varepsilon)=0 zero we obtain

    b03=23[b12b21(a21+b12)p12],q03=13[2b21(1+2p21)p12+2(b21p12)q12].\displaystyle b_{03}=\frac{2}{3}\big{[}b_{12}b_{21}-(a_{21}+b_{12})p_{12}\big{]},\quad q_{03}=\frac{1}{3}\big{[}2b_{21}-(1+2p_{21})p_{12}+2(b_{21}-p_{12})q_{12}\big{]}.

    Then, we have the 44th generalized Lyapunov constant,

    V4(ε)=\displaystyle V_{4}(\varepsilon)= 1645ε3[a02+(p02p12)ε2]{2(a21+b12)(b21+2p12)\displaystyle\frac{16}{45}\varepsilon^{3}\big{[}a_{02}+(p_{02}-p_{12})\varepsilon^{2}\big{]}\big{\{}2(a_{21}+b_{12})(b_{21}+2p_{12})
    +[3b21+2p21(b21+2p12)+2q12(b21+2p12)]ε2}.\displaystyle+\big{[}3b_{21}+2p_{21}(b_{21}+2p_{12})+2q_{12}(b_{21}+2p_{12})\big{]}\varepsilon^{2}\big{\}}.

    Two subcases follow the ε5\varepsilon^{5}-order term in V4(ε)V_{4}(\varepsilon) to be zero: (i-1-1-1) p12=b212p_{12}=-\frac{b_{21}}{2}, and (i-1-1-2) p12b212p_{12}\neq-\frac{b_{21}}{2}.

    (i-1-1-1) If p12=b212p_{12}=-\frac{b_{21}}{2}, we obtain b21=0b_{21}=0 by V4(ε)=0V_{4}(\varepsilon)=0, giving the condition V.

    (i-1-1-2) If p12b212p_{12}\neq-\frac{b_{21}}{2}, we obtain b12=a21b_{12}=-a_{21} and

    p21=3b212(b21+2p12)q122(b21+2p12)p_{21}=\frac{-3b_{21}-2(b_{21}+2p_{12})q_{12}}{2(b_{21}+2p_{12})}

    from V4(ε)=0V_{4}(\varepsilon)=0. Further, we have

    V5(ε)=\displaystyle V_{5}(\varepsilon)= π144(b21+2p12)3b21ε5[45a022(b21+2p12)3+2(b21+2p12)2V5aε2\displaystyle-\frac{\pi}{144(b_{21}+2p_{12})^{3}}\,b_{21}\varepsilon^{5}\big{[}45a_{02}^{2}(b_{21}+2p_{12})^{3}+2(b_{21}+2p_{12})^{2}V_{5a}\varepsilon^{2}
    +(b21+2p12)V5bε418V5cε6],\displaystyle\hskip 119.24506pt+(b_{21}+2p_{12})V_{5b}\varepsilon^{4}-18V_{5c}\varepsilon^{6}\big{]},

    where

    V5a=\displaystyle V_{5a}= 45a02(p02p12)(b21+2p12)+9a03(2b21+p12)\displaystyle\ 45a_{02}(p_{02}-p_{12})(b_{21}+2p_{12})+9a_{03}(2b_{21}+p_{12})
    +2a21(p122b21)(9a21+4b212)+80a21b21p122,\displaystyle+2a_{21}(p_{12}-2b_{21})(9a_{21}+4b_{21}^{2})+80a_{21}b_{21}p_{12}^{2},
    V5b=\displaystyle V_{5b}= b212(108a21+32b212+45p022)+18b21(2b21+5p12)p03\displaystyle\ b_{21}^{2}(108a_{21}+32b_{21}^{2}+45p_{02}^{2})+18b_{21}(2b_{21}+5p_{12})p_{03}
    2b21(9a2128b212+45b21p0290p022)p12\displaystyle-2b_{21}(9a_{21}-28b_{21}^{2}+45b_{21}p_{02}-90p_{02}^{2})p_{12}
    +3[96a2129b212120(b21+p12)p02+60p022+12p03]p122212b21p123\displaystyle+3\big{[}96a_{21}-29b_{21}^{2}-120(b_{21}+p_{12})p_{02}+60p_{02}^{2}+12p_{03}\big{]}p_{12}^{2}-212b_{21}p_{12}^{3}
    140p124+8(2p12+b21)[(2b21p12)(9a21+2b212)20b21p122]q12,\displaystyle-140p_{12}^{4}+8(2p_{12}+b_{21})\big{[}(2b_{21}-p_{12})(9a_{21}+2b_{21}^{2})-20b_{21}p_{12}^{2}\big{]}q_{12},
    V5c=\displaystyle V_{5c}= [b214p12+2(b21+2p12)q12][2b21(b21+2p12)+2(b212p122)q12\displaystyle\ \big{[}b_{21}-4p_{12}+2(b_{21}+2p_{12})q_{12}\big{]}\big{[}2b_{21}(b_{21}+2p_{12})+2(b_{21}^{2}-p_{12}^{2})q_{12}
    +3(2p12+b21q12)p12].\displaystyle+3(2p_{12}+b_{21}q_{12})p_{12}\big{]}.

    It follows from V5(ε)=0V_{5}(\varepsilon)=0 that b21=0b_{21}=0, which is included in the condition V. Otherwise, the ε5\varepsilon^{5}-order term in V5(ε)V_{5}(\varepsilon) is non-zero.

    (i-1-2) Assume q210q_{21}\neq 0 and q12=1q_{12}=-1. Letting the ε3\varepsilon^{3}-order and ε5\varepsilon^{5}-order terms in V3(ε)V_{3}(\varepsilon) zero, we obtain

    b03=23[b12b21(a21+b12)p12],q03=13(p122p12p21+2b12q21).\displaystyle b_{03}=\frac{2}{3}\big{[}b_{12}b_{21}-(a_{21}+b_{12})p_{12}\big{]},\quad q_{03}=\frac{1}{3}(p_{12}-2p_{12}p_{21}+2b_{12}q_{21}).

    Then, we have the 44th generalized Lyapunov constant,

    V4(ε)=\displaystyle V_{4}(\varepsilon)= 1645ε3[a02+(p02p12)ε2]{2(a21+b12)(b21+2p12)\displaystyle\frac{16}{45}\varepsilon^{3}\big{[}a_{02}+(p_{02}-p_{12})\varepsilon^{2}\big{]}\big{\{}2(a_{21}+b_{12})(b_{21}+2p_{12})
    +[b214p12+2(b21+2p12)p21+2(a21+b12)q21]ε2+(1+2p21)q21ε4}.\displaystyle+\big{[}b_{21}-4p_{12}+2(b_{21}+2p_{12})p_{21}+2(a_{21}+b_{12})q_{21}\big{]}\varepsilon^{2}+(1+2p_{21})q_{21}\varepsilon^{4}\big{\}}.

    Letting the ε9\varepsilon^{9}-order terms in V4(ε)V_{4}(\varepsilon) zero we have p21=12p_{21}=-\frac{1}{2}.

    (i-1-2-1) If b12=a21b_{12}=-a_{21}, the ε3\varepsilon^{3}-order term in V4(ε)V_{4}(\varepsilon) becomes zero. By using V4(ε)=0V_{4}(\varepsilon)=0 we have p12=0p_{12}=0. Then, the 55th generalized Lyapunov constant is simplified as

    V5(ε)=\displaystyle V_{5}(\varepsilon)= π144ε5(b21+q21ε2)[45a022+2(18a0336a21216a21b212+45a02p02)ε2\displaystyle\ -\frac{\pi}{144}\varepsilon^{5}(b_{21}+q_{21}\varepsilon^{2})\big{[}45a_{02}^{2}+2(18a_{03}-36a_{21}^{2}-16a_{21}b_{21}^{2}+45a_{02}p_{02})\varepsilon^{2}
    (36a2145p02236p03+64a21b21q21)ε432a21q212ε6].\displaystyle-(36a_{21}-45p_{02}^{2}-36p_{03}+64a_{21}b_{21}q_{21})\varepsilon^{4}-32a_{21}q_{21}^{2}\varepsilon^{6}\big{]}.

    If b21=0b_{21}=0, we obtain the condition included in the condition V. Otherwise, the ε5\varepsilon^{5}-order term in V5(ε)V_{5}(\varepsilon) does not equal zero.

    (i-1-2-2) If p12=b212p_{12}=-\frac{b_{21}}{2}, the ε3\varepsilon^{3}-order term in V4(ε)V_{4}(\varepsilon) again becomes zero. From V4(ε)=0V_{4}(\varepsilon)=0 we obtain b12=3b212a21q212q21b_{12}=\frac{-3b_{21}-2a_{21}q_{21}}{2q_{21}}. Then, we have the 55th generalized Lyapunov constant,

    V5(ε)=\displaystyle V_{5}(\varepsilon)= π576q212ε5(b21+q21ε2)[V~5a+18V~5bε2+q21V~5cε4+16q213(13b21+8a21q21)ε6],\displaystyle\frac{\pi}{576q_{21}^{2}}\varepsilon^{5}(b_{21}+q_{21}\varepsilon^{2})\big{[}\widetilde{V}_{5a}+18\widetilde{V}_{5b}\varepsilon^{2}+q_{21}\widetilde{V}_{5c}\varepsilon^{4}+16q_{21}^{3}(13b_{21}+8a_{21}q_{21})\varepsilon^{6}\big{]},

    where

    V~5a=\displaystyle\widetilde{V}_{5a}= 36[9a21b212b21(3a0310a212)q215a022q212],\displaystyle\ 36\big{[}9a_{21}b_{21}^{2}-b_{21}(3a_{03}-10a_{21}^{2})q_{21}-5a_{02}^{2}q_{21}^{2}\big{]},
    V~5b=\displaystyle\widetilde{V}_{5b}= 9b212+2b21(9b212+19a213p03)q21\displaystyle\ 9b_{21}^{2}+2b_{21}(9b_{21}^{2}+19a_{21}-3p_{03})q_{21}
    2[4a03+5(b21+2p02)a028a21(a21+b212)]q212,\displaystyle-2\big{[}4a_{03}+5(b_{21}+2p_{02})a_{02}-8a_{21}(a_{21}+b_{21}^{2})\big{]}q_{21}^{2},
    V~5c=\displaystyle\widetilde{V}_{5c}= 252b21+9[16a21+63b21220(b21+p02)p0216p03]q21+416a21b21q212.\displaystyle\ 252b_{21}+9\big{[}16a_{21}+63b_{21}^{2}-20(b_{21}+p_{02})p_{02}-16p_{03}\big{]}q_{21}+416a_{21}b_{21}q_{21}^{2}.

    If b21=0b_{21}=0 we have V~5a=180a022q2120\widetilde{V}_{5a}=-180a_{02}^{2}q_{21}^{2}\neq 0. Hence, by using V5(ε)=0V_{5}(\varepsilon)=0 we have a21=138q21b210a_{21}=-\frac{13}{8q_{21}}b_{21}\neq 0, V~5b=0\widetilde{V}_{5b}=0 and

    a03=\displaystyle a_{03}= 196b21q21(377b213160a022q212),\displaystyle\ \frac{1}{96b_{21}q_{21}}(377b_{21}^{3}-160a_{02}^{2}q_{21}^{2}),
    p03=\displaystyle p_{03}= 1144q21[b21(18109b21q21)180(b21+p02)p02q21].\displaystyle\ \frac{1}{144q_{21}}\big{[}b_{21}(18-109b_{21}q_{21})-180(b_{21}+p_{02})p_{02}q_{21}\big{]}.

    Setting the ε17\varepsilon^{17}-order term in V6(ε)V_{6}(\varepsilon) zero we have p02=12b21p_{02}=-\frac{1}{2}b_{21}. Further, we have the simplified Lyapunov constant,

    V6(ε)=\displaystyle V_{6}(\varepsilon)= 19450b212q212ε7(b21+q21ε2)[96a02q21(7567b213+296b214q212360a022q213)\displaystyle\frac{-1}{9450b_{21}^{2}q_{21}^{2}}\varepsilon^{7}(b_{21}+q_{21}\varepsilon^{2})\big{[}96a_{02}q_{21}(7567b_{21}^{3}+296b_{21}^{4}q_{21}-2360a_{02}^{2}q_{21}^{3})
    22050πb21(16b213+2b214q215a022q213)ε+1152a02b212q212(3578b21q21)ε2\displaystyle-22050\pi b_{21}(16b_{21}^{3}+2b_{21}^{4}q_{21}-5a_{02}^{2}q_{21}^{3})\varepsilon+1152a_{02}b_{21}^{2}q_{21}^{2}(35-78b_{21}q_{21})\varepsilon^{2}
    11025πq21(16b213+2b214q215a022q213)ε318432a02b21q213(1+3b21q21)ε4\displaystyle-11025\pi q_{21}(16b_{21}^{3}+2b_{21}^{4}q_{21}-5a_{02}^{2}q_{21}^{3})\varepsilon^{3}-18432a_{02}b_{21}q_{21}^{3}(1+3b_{21}q_{21})\varepsilon^{4}
    16384a02b21q215ε6],\displaystyle-16384a_{02}b_{21}q_{21}^{5}\varepsilon^{6}\big{]},

    which has a non-zero ε15\varepsilon^{15}-order term.

    (i-2) If a120a_{12}\neq 0, we obtain the 33rd generalized Lyapunov constant,

    V3(ε)=π4ε{2a12(a21+b12)+[a12(1+2p21)+3b032b12b21+2(a21+b12)p12]ε2\displaystyle V_{3}(\varepsilon)=\frac{\pi}{4}\varepsilon\big{\{}2a_{12}(a_{21}+b_{12})+\big{[}a_{12}(1+2p_{21})+3b_{03}-2b_{12}b_{21}+2(a_{21}+b_{12})p_{12}\big{]}\varepsilon^{2}
    +[3q032b21(1+q12)+(1+2p21+2q12)p122b12q21]ε42(1+q12)q21ε6}.\displaystyle+\big{[}3q_{03}-2b_{21}(1+q_{12})+(1+2p_{21}+2q_{12})p_{12}-2b_{12}q_{21}\big{]}\varepsilon^{4}-2(1+q_{12})q_{21}\varepsilon^{6}\big{\}}.

    Similar to the previous case, we have the following two subcases: (i-2-1) q21=0q_{21}=0 and (i-2-2) q12=1q_{12}=-1.

    (i-2-1) If q21=0q_{21}=0, setting the ε\varepsilon-order, ε3\varepsilon^{3}-order and ε5\varepsilon^{5}-order terms in V3(ε)V_{3}(\varepsilon) to equal zero yields

    b12=\displaystyle b_{12}= a21,\displaystyle-a_{21},
    b03=\displaystyle b_{03}= 13[a12(1+2p21+2q12)+2a21b21],\displaystyle-\frac{1}{3}\big{[}a_{12}(1+2p_{21}+2q_{12})+2a_{21}b_{21}\big{]},
    q03=\displaystyle q_{03}= 13[2b21(1+q12)(1+2p21+2q12)p12].\displaystyle\frac{1}{3}\big{[}2b_{21}(1+q_{12})-(1+2p_{21}+2q_{12})p_{12}\big{]}.

    Further, we have

    V4(ε)=\displaystyle V_{4}(\varepsilon)= 1645ε3[a02+(p02p12)ε2]{4a12(p21+q12)+[3b21+2(b21+2p12)(p21+q12)]ε2}.\displaystyle\frac{16}{45}\varepsilon^{3}\big{[}a_{02}+(p_{02}-p_{12})\varepsilon^{2}\big{]}\big{\{}4a_{12}(p_{21}+q_{12})+\big{[}3b_{21}+2(b_{21}+2p_{12})(p_{21}+q_{12})\big{]}\varepsilon^{2}\big{\}}.

    By using V4(ε)=0V_{4}(\varepsilon)=0 we have b21=0b_{21}=0 and q12=p21q_{12}=-p_{21}, giving the first condition in VI.

    (i-2-2) If q12=1q_{12}=-1, vanishing each ε\varepsilon order term in V3(ε)V_{3}(\varepsilon) yields

    b12=a21,b03=13(a122a21b212a12p21),q03=13(p122p12p212a21q21).\displaystyle b_{12}=-a_{21},\quad b_{03}=\frac{1}{3}(a_{12}-2a_{21}b_{21}-2a_{12}p_{21}),\quad q_{03}=\frac{1}{3}(p_{12}-2p_{12}p_{21}-2a_{21}q_{21}).

    Then, we obtain the 44th generalized Lyapunov constant,

    V4(ε)=\displaystyle V_{4}(\varepsilon)= 1645ε3[a02+(p02p12)ε2]\displaystyle\frac{16}{45}\varepsilon^{3}\big{[}a_{02}+(p_{02}-p_{12})\varepsilon^{2}\big{]}
    ×{4a12(p211)+[b21(1+2p21)+4(p211)p12]ε2q21(1+2p21)ε4}.\displaystyle\times\big{\{}4a_{12}(p_{21}-1)+\big{[}b_{21}(1+2p_{21})+4(p_{21}-1)p_{12}\big{]}\varepsilon^{2}-q_{21}(1+2p_{21})\varepsilon^{4}\big{\}}.

    It follows from V4(ε)=0V_{4}(\varepsilon)=0 that b21=0b_{21}=0, p21=1p_{21}=1 and q21=0q_{21}=0, which also leads to the first condition in VI.

  2. (ii)

    If a02=a12<0a_{02}=a_{12}<0, similar to the analysis for the subcase (i-2), replacing a02a_{02} by a12a_{12} in V4(ε)V_{4}(\varepsilon), then we obtain the second condition in VI.

The above discussions (or see Table 2) show that the conditions V and VI are necessary for the origin of the perturbed system (37) to be a center, and so they are necessary conditions for (±1\pm 1,0) of the unperturbed system (9) to be nilpotent centers. We can also prove that these conditions are sufficient.

If the condition V in (15) holds, system (9) is reduced to

(x˙y˙)={(a21y+a21x2y+a02y2+a03y3,x2+x32+b12xy2,),ify>0,(a21y+a21x2ya02y2+a03y3,x2+x32+b12xy2,),ify<0.\left(\begin{array}[]{cc}\dot{x}\\ \dot{y}\end{array}\right)=\left\{\begin{aligned} &\left(\begin{aligned} &-a_{21}y+a_{21}x^{2}y+a_{02}y^{2}+a_{03}y^{3},\\[2.15277pt] &-\tfrac{x}{2}+\tfrac{x^{3}}{2}+b_{12}xy^{2},\end{aligned}\right),&\text{if}\ \ y>0,\\[4.30554pt] &\left(\begin{aligned} &-a_{21}y+a_{21}x^{2}y-a_{02}y^{2}+a_{03}y^{3},\\[2.15277pt] &-\tfrac{x}{2}+\tfrac{x^{3}}{2}+b_{12}xy^{2},\end{aligned}\right),&\text{if}\ \ y<0.\end{aligned}\right. (38)

It is easy to see that system (38) is symmetric with respect to the xx-axis, and so by the symmetry of switching systems redefined in Theorem 2.1 of [35], it implies that the singular points (±1\pm 1,0) of the system (9) are bi-center, which consists of two 2nd-order nilpotent cusps.

If the condition VI in (15) holds, via xx+1x\rightarrow x+1, system (9) becomes

(x˙y˙)={(2a21xy+(a02+a12)y2+a21x2y3b03xy2+a03y3,x+32x2a21y2+12x3a21xy2+b03y3,),ify>0,(2a21xy+a21x2y(a02+a12+6b03)y23b03xy2+a03y3,x+32x2a21y2+12x3a21xy2+b03y3,),ify<0.\left(\begin{array}[]{cc}\dot{x}\\ \dot{y}\end{array}\right)=\left\{\begin{aligned} &\left(\begin{aligned} &2a_{21}xy+(a_{02}+a_{12})y^{2}+a_{21}x^{2}y-3b_{03}xy^{2}+a_{03}y^{3},\\[2.15277pt] &x+\tfrac{3}{2}x^{2}-a_{21}y^{2}+\tfrac{1}{2}x^{3}-a_{21}xy^{2}+b_{03}y^{3},\end{aligned}\right),&\text{if}\ \ y>0,\\[4.30554pt] &\left(\begin{aligned} &2a_{21}xy+a_{21}x^{2}y-(a_{02}+a_{12}+6b_{03})y^{2}-3b_{03}xy^{2}+a_{03}y^{3},\\[2.15277pt] &x+\tfrac{3}{2}x^{2}-a_{21}y^{2}+\tfrac{1}{2}x^{3}-a_{21}xy^{2}+b_{03}y^{3},\end{aligned}\right),&\text{if}\ \ y<0.\end{aligned}\right. (39)

Then it can be shown that the upper and lower systems in (39) are Hamiltonian systems, having respectively the Hamiltonian functions:

H+(x,y)\displaystyle H^{+}(x,y) =12x212x3+a21xy2+a02+a123y318x4+a212x2y2b03xy3+a034y4,\displaystyle=-\frac{1}{2}x^{2}-\frac{1}{2}x^{3}+a_{21}xy^{2}+\frac{a_{02}+a_{12}}{3}y^{3}-\frac{1}{8}x^{4}+\frac{a_{21}}{2}x^{2}y^{2}-b_{03}xy^{3}+\frac{a_{03}}{4}y^{4}, (40)
H(x,y)\displaystyle H^{-}(x,y) =12x212x3+a21xy2a02+a12+6b033y318x4+a212x2y2b03xy3+a034y4.\displaystyle=-\frac{1}{2}x^{2}\!-\!\frac{1}{2}x^{3}\!+\!a_{21}xy^{2}\!-\!\frac{a_{02}\!+\!a_{12}\!+\!6b_{03}}{3}y^{3}\!-\!\frac{1}{8}x^{4}\!+\!\frac{a_{21}}{2}x^{2}y^{2}\!-\!b_{03}xy^{3}\!+\!\frac{a_{03}}{4}y^{4}.

It is straightforward to verify that the condition H+(x,0)H(x,0)H^{+}(x,0)\equiv H^{-}(x,0) in Proposition 2.1 of [10] is satisfied, indicating that the origin of system (39) is a center. Hence, system (9) has bi-center at (±1\pm 1,0).

Example 4.4.

The global phase portraits of system (9) corresponding to the bi-center conditions V and VI, with two sets of parameter values, as given in Figure 5, show the case of bi-center at (±1,0)(\pm 1,0).

Refer to caption
Refer to caption
Figure 5: The phase portraits of system (9) showing bi-center at (±1,0)(\pm 1,0) for (a) Condition V: a02=1,a12=b03=b02=b21=0,a21=a03=b12=1a_{02}=-1,\,a_{12}=b_{03}=b_{02}=b_{21}=0,\,a_{21}=a_{03}=b_{12}=1; and (b) Condition VI: a02=4,a21=b03=1,a12=3,a03=b12=1,b02=b21=0a_{02}=-4,\,a_{21}=b_{03}=-1,\,a_{12}=3,\,a_{03}=b_{12}=1,\,b_{02}=b_{21}=0.

The proof for Theorem 2.2 is completed.

5 The proof of Theorem 2.3

In this section, we will perturb system (9) with the 66 center conditions in (15) to find the maximal number of small-amplitude limit cycles bifurcating from the bi-center. Since system (9) is symmetric with the origin, we only need to study the limit cycles around the singular point (1,0)(1,0). Using each of the 66 center conditions, we can show that system (9) can have 99 small-amplitude limit cycles around each of the bi-center, and then plus additional 11 small limit cycle from the pseudo-Hopf bifurcation, leading to a total of 2020 small-amplitude limit cycles bifurcating from the bi-center. Perturbations up to cubic terms are applied to system (9), and it will be shown that there exist maximal 99 (including both system parameters and perturbation parameters) independent perturbation parameters. Therefore, 99 small-amplitude limit cycles may bifurcate from each of the bi-center. Since the proofs for the 66 center conditions I-VI are similar, we only prove one case for the center condition VI. It has been noted that for the first 55 center conditions I-V, when the conditions are obtained to satisfy V1=V2==V8=0V_{1}=V_{2}=\cdots=V_{8}=0, then V9V_{9} and V10V_{10} equal zero simultaneously under a same condition, for which V110V_{11}\neq 0, implying that maximal 99 small-amplitude limit cycles many exist around each of the bi-center. For the 66th center condition VI, solutions exist such that V1=V2==V9=0V_{1}=V_{2}=\cdots=V_{9}=0, V100V_{10}\neq 0, again indicating the existence of maximal 99 small-amplitude limit cycles.

In order to include all possible perturbations, we add the perturbations including all εk\varepsilon^{k} (k1k\geq 1) order terms to system (9) under the center condition VI to obtain

(x˙y˙)=((a21y+a02y2+a21x2y3b03xy2+a03y3δ12ε(xx3)+P+(x,y,ε),12x+12x3a21xy2+b03y3+b2ε3(xx2)+δ1εy+Q+(x,y,ε),),ify>0,(a21ya02y2+a21x2y3b03xy2+a03y3δ12ε(xx3)+P(x,y,ε),12x+12x3a21xy2+b03y3+b2ε3(x+x2)+δ1εy+Q(x,y,ε),),ify<0,{\small\left(\begin{array}[]{cc}\dot{x}\\ \dot{y}\end{array}\right)=\left(\begin{aligned} &\left(\begin{aligned} &-a_{21}y+a_{02}y^{2}+a_{21}x^{2}y-3b_{03}xy^{2}+a_{03}y^{3}-\tfrac{\delta_{1}}{2}\varepsilon(x-x^{3})\\ &+P^{+}(x,y,\varepsilon),\\[1.93748pt] &-\tfrac{1}{2}x+\tfrac{1}{2}x^{3}-a_{21}xy^{2}+b_{03}y^{3}+\tfrac{b}{2}\varepsilon^{3}(x-x^{2})+\delta_{1}\varepsilon y+Q^{+}(x,y,\varepsilon),\end{aligned}\right),&\text{if}\ \ y>0,\\[3.87498pt] &\left(\begin{aligned} &-a_{21}y-a_{02}y^{2}+a_{21}x^{2}y-3b_{03}xy^{2}+a_{03}y^{3}-\tfrac{\delta_{1}}{2}\varepsilon(x-x^{3})\\ &+P^{-}(x,y,\varepsilon),\\[1.93748pt] &-\tfrac{1}{2}x+\tfrac{1}{2}x^{3}-a_{21}xy^{2}+b_{03}y^{3}+\tfrac{b}{2}\varepsilon^{3}(x+x^{2})+\delta_{1}\varepsilon y+Q^{-}(x,y,\varepsilon),\end{aligned}\right),&\text{if}\ \ y<0,\end{aligned}\right.} (41)

where

P+(x,y,ε)=k=1i+j=03εkpijkxiyj,P(x,y,ε)=k=1i+j=03εk(1)i+j+1pijkxiyj,\displaystyle P^{+}(x,y,\varepsilon)=\sum_{k=1}\sum_{i+j=0}^{3}\varepsilon^{k}p_{ijk}x^{i}y^{j},\qquad P^{-}(x,y,\varepsilon)=\sum_{k=1}\sum_{i+j=0}^{3}\varepsilon^{k}(-1)^{i+j+1}p_{ijk}x^{i}y^{j}, (42)
Q+(x,y,ε)=k=1i+j=03εkqijkxiyj,Q(x,y,ε)=k=1i+j=03εk(1)i+j+1qijkxiyj.\displaystyle Q^{+}(x,y,\varepsilon)=\sum_{k=1}\sum_{i+j=0}^{3}\varepsilon^{k}q_{ijk}x^{i}y^{j},\qquad Q^{-}(x,y,\varepsilon)=\sum_{k=1}\sum_{i+j=0}^{3}\varepsilon^{k}(-1)^{i+j+1}q_{ijk}x^{i}y^{j}.

With the transformation xx+1x\rightarrow x+1, the first two polynomials in (42) become

p+(x,y,ε)=\displaystyle p^{+}(x,y,\varepsilon)= k=1[p00k+p10k+p20k+p30k+(p10k+2p20k+3p30k)x\displaystyle\,\sum_{k=1}\big{[}p_{00k}+p_{10k}+p_{20k}+p_{30k}+(p_{10k}+2p_{20k}+3p_{30k})x (43)
+(p20k+3p30k)x2+p30kx3+(p01k+p11k+p21k)y\displaystyle\qquad+(p_{20k}+3p_{30k})x^{2}+p_{30k}x^{3}+(p_{01k}+p_{11k}+p_{21k})y
+(p11k+2p21k)xy+p21kx2y+(p02k+p12k)y2+p12kxy2+p03ky3]εk,\displaystyle\qquad+(p_{11k}+2p_{21k})xy+p_{21k}x^{2}y+(p_{02k}+p_{12k})y^{2}+p_{12k}xy^{2}+p_{03k}y^{3}\big{]}\varepsilon^{k},
p(x,y,ε)=\displaystyle p^{-}(x,y,\varepsilon)= k=1[p00k+p10kp20k+p30k+(p10k2p20k+3p30k)x\displaystyle\sum_{k=1}\big{[}-p_{00k}+p_{10k}-p_{20k}+p_{30k}+(p_{10k}-2p_{20k}+3p_{30k})x
(p20k3p30k)x2+p30kx3+(p01kp11k+p21k)y\displaystyle\qquad-(p_{20k}-3p_{30k})x^{2}+p_{30k}x^{3}+(p_{01k}-p_{11k}+p_{21k})y
(p11k2p21k)xy+p21kx2y(p02kp12k)y2+p12kxy2+p03ky3]εk.\displaystyle\qquad-(p_{11k}-2p_{21k})xy+p_{21k}x^{2}y-(p_{02k}-p_{12k})y^{2}+p_{12k}xy^{2}+p_{03k}y^{3}\big{]}\varepsilon^{k}.

In order to keep the origin of the corresponding shifted system to be a monodromic point, we eliminate the constants and the xx and yy terms in each εk\varepsilon^{k}-order term of p+(x,y,ε)p^{+}(x,y,\varepsilon) and p(x,y,ε)p^{-}(x,y,\varepsilon), and obtain

p00k=p10k=p20k=p11k=p30k=0,p01k=p21k,k1.\displaystyle p_{00k}=p_{10k}=p_{20k}=p_{11k}=p_{30k}=0,\quad p_{01k}=-p_{21k},\quad\forall k\geq 1.

Similarly, simplifying Q±(x,y,ε)Q^{\pm}(x,y,\varepsilon) in (42) yields

q00k=q10k=q20k=q11k=q30k=0,q01k=q21k,k1.\displaystyle q_{00k}=q_{10k}=q_{20k}=q_{11k}=q_{30k}=0,\quad q_{01k}=-q_{21k},\quad\forall k\geq 1.

Next, with a simple parametrization: p02k+p12kp02kp_{02k}+p_{12k}\rightarrow p_{02k} and q02k+q12kq02kq_{02k}+q_{12k}\rightarrow q_{02k}, we have the following four translational perturbations:

p+(x,y,ε)=\displaystyle p^{+}(x,y,\varepsilon)= k=1(p30kx3+2p21kxy+p21kx2y+p02ky2+p12kxy2+p03ky3)εk,\displaystyle\ \sum_{k=1}(p_{30k}x^{3}+2p_{21k}xy+p_{21k}x^{2}y+p_{02k}y^{2}+p_{12k}xy^{2}+p_{03k}y^{3})\varepsilon^{k}, (44)
p(x,y,ε)=\displaystyle p^{-}(x,y,\varepsilon)= k=1[p30kx3+2p21kxy+p21kx2y+(2p12kp02k)y2+p12kxy2+p03ky3]εk,\displaystyle\ \sum_{k=1}\big{[}p_{30k}x^{3}+2p_{21k}xy+p_{21k}x^{2}y+(2p_{12k}-p_{02k})y^{2}+p_{12k}xy^{2}+p_{03k}y^{3}\big{]}\varepsilon^{k},
q+(x,y,ε)=\displaystyle q^{+}(x,y,\varepsilon)= k=1(q30kx3+2q21kxy+q21kx2y+q02ky2+q12kxy2+q03ky3)εk,\displaystyle\ \sum_{k=1}(q_{30k}x^{3}+2q_{21k}xy+q_{21k}x^{2}y+q_{02k}y^{2}+q_{12k}xy^{2}+q_{03k}y^{3})\varepsilon^{k},
q(x,y,ε)=\displaystyle q^{-}(x,y,\varepsilon)= k=1[q30kx3+2q21kxy+q21kx2y+(2q12kq02k)y2+q12kxy2+q03ky3]εk.\displaystyle\ \sum_{k=1}\big{[}q_{30k}x^{3}+2q_{21k}xy+q_{21k}x^{2}y+(2q_{12k}-q_{02k})y^{2}+q_{12k}xy^{2}+q_{03k}y^{3}]\varepsilon^{k}.

Therefore, under the transformation xx+1x\rightarrow x+1, system (41) becomes the following perturbed system,

(x˙y˙)={(2a21xy+(a023b03)y2+a21x2y3b03xy2+a03y3+δ1ε(x+32x2+12x3)ε2y+p+(x,y,ε),x+32x2+12x3a21y2a21xy2+b03y3+δ1εyb2ε3(x+x2)+q+(x,y,ε),),ify>0,(2a21xy(a02+3b03)y2+a21x2y3b03xy2+a03y3+δ1ε(x+32x2+12x3)ε2y+p(x,y,ε),x+32x2+12x3a21y2a21xy2+b03y3+δ1εy+b2ε3(2+3x+x2)+q(x,y,ε),),ify<0.{\small\left(\begin{array}[]{cc}\dot{x}\\ \dot{y}\end{array}\right)=\left\{\begin{aligned} &\left(\begin{aligned} &2a_{21}xy+(a_{02}-3b_{03})y^{2}+a_{21}x^{2}y-3b_{03}xy^{2}+a_{03}y^{3}\\ &+\delta_{1}\varepsilon(x+\tfrac{3}{2}x^{2}+\tfrac{1}{2}x^{3})-\varepsilon^{2}y+p^{+}(x,y,\varepsilon),\\[1.93748pt] &x+\tfrac{3}{2}x^{2}+\tfrac{1}{2}x^{3}-a_{21}y^{2}-a_{21}xy^{2}+b_{03}y^{3}\\ &+\delta_{1}\varepsilon y-\tfrac{b}{2}\varepsilon^{3}(x+x^{2})+q^{+}(x,y,\varepsilon),\end{aligned}\right),&\text{if}\ \ y>0,\\[3.87498pt] &\left(\begin{aligned} &2a_{21}xy-(a_{02}+3b_{03})y^{2}+a_{21}x^{2}y-3b_{03}xy^{2}+a_{03}y^{3}\\ &+\delta_{1}\varepsilon(x+\tfrac{3}{2}x^{2}+\tfrac{1}{2}x^{3})-\varepsilon^{2}y+p^{-}(x,y,\varepsilon),\\[1.93748pt] &x+\tfrac{3}{2}x^{2}+\tfrac{1}{2}x^{3}-a_{21}y^{2}-a_{21}xy^{2}+b_{03}y^{3}\\ &+\delta_{1}\varepsilon y+\tfrac{b}{2}\varepsilon^{3}(2+3x+x^{2})+q^{-}(x,y,\varepsilon),\end{aligned}\right),&\text{if}\ \ y<0.\end{aligned}\right.} (45)

By the approach described in [43] we introduce the following near-identity state transformation and time rescaling:

xx+d1(ε)x+d2(ε)y,yy+d3(ε)x+d4(ε)y,tt+d5(ε)t,\begin{array}[]{ll}x\rightarrow x+d_{1}(\varepsilon)x+d_{2}(\varepsilon)y,\quad y\rightarrow y+d_{3}(\varepsilon)x+d_{4}(\varepsilon)y,\quad t\rightarrow t+d_{5}(\varepsilon)t,\\ \end{array} (46)

for the upper system in (45), where

di(ε)=di1ε+di2ε2++dinεn,i=1,2,,5.d_{i}(\varepsilon)=d_{i1}\varepsilon+d_{i2}\varepsilon^{2}+\cdots+d_{in}\varepsilon^{n},\quad i=1,2,\cdots,5. (47)

A similar transformation can be used to simplify the lower system in (45). Note that the unperturbed system of (45) is unchanged by the identity map (46)|ε=0|_{\varepsilon=0}. Therefore, we may obtain proper di(ε)d_{i}(\varepsilon)’s to simplify the perturbations without loss of generality. By substituting (46) into system (45) and taking the ε\varepsilon-order terms, we obtain

x˙=d21x12(3d214a21d31)x2+2(a02d313b03d31+a21d41+a21d51+p211)xy(a02d113b03d113a21d212a02d41+6b03d41a02d51+3b03d51p021)y212(d212a21d31)x3+(a21d116b03d31+a21d41+a21d51+p211)x2y+(3a21d21+3a03d316b03d413b03d51+p121)xy2(a03d11+4b03d213a03d41a03d51p031)y3,\begin{array}[]{ll}\dot{x}=\!\!\!&-d_{21}x-\frac{1}{2}(3d_{21}-4a_{21}d_{31})x^{2}+2(a_{02}d_{31}-3b_{03}d_{31}+a_{21}d_{41}+a_{21}d_{51}+p_{211})xy\\[2.15277pt] &-(a_{02}d_{11}-3b_{03}d_{11}-3a_{21}d_{21}-2a_{02}d_{41}+6b_{03}d_{41}-a_{02}d_{51}+3b_{03}d_{51}-p_{021})y^{2}\\[2.15277pt] &-\frac{1}{2}(d_{21}-2a_{21}d_{31})x^{3}+(a_{21}d_{11}-6b_{03}d_{31}+a_{21}d_{41}+a_{21}d_{51}+p_{211})x^{2}y\\[2.15277pt] &+(3a_{21}d_{21}+3a_{03}d_{31}-6b_{03}d_{41}-3b_{03}d_{51}+p_{121})xy^{2}\\[2.15277pt] &-(a_{03}d_{11}+4b_{03}d_{21}-3a_{03}d_{41}-a_{03}d_{51}-p_{031})y^{3},\end{array} (48)

and

y˙=(d11d41+d51)x+d21y+32(2d11d41+d51)x2+(3d214a21d31+2q211)xy(a02d313b03d31+a21d41+a21d51q021)y2+12(3d216a21d31+2q211)x2y(a21d116b03d31+a21d41+a21d51q121)xy2(a21d21+a03d312b03d41b03d51q031)y3+12(3d11d41+d51)x3.\begin{array}[]{ll}\dot{y}=\!\!\!&(d_{11}-d_{41}+d_{51})x+d_{21}\,y+\frac{3}{2}(2d_{11}-d_{41}+d_{51})x^{2}+(3d_{21}-4a_{21}d_{31}+2q_{211})xy\\[2.15277pt] &-(a_{02}d_{31}-3b_{03}d_{31}+a_{21}d_{41}+a_{21}d_{51}-q_{021})\,y^{2}+\frac{1}{2}(3d_{21}-6a_{21}d_{31}+2q_{211})x^{2}y\\[2.15277pt] &-(a_{21}d_{11}-6b_{03}d_{31}+a_{21}d_{41}+a_{21}d_{51}-q_{121})xy^{2}\\[2.15277pt] &-(a_{21}d_{21}+a_{03}d_{31}-2b_{03}d_{41}-b_{03}d_{51}-q_{031})y^{3}+\frac{1}{2}(3d_{11}-d_{41}+d_{51})x^{3}.\end{array} (49)

Now, simplify setting

d11=12(a023b03)(3b032a03a21){a02a21p031b03[3a21p0316b03(p021p121)2a02p121]a03[2a21p021+(a023b03)q121]},d21=0,d31=16a03(3b032a03a21)(3a21b03p031+2a03a21p1213a03b03q121),d41=1a03(3b03a02)(a03p021a02p031+3b03p031),d51=12a03(a023b03)(3b032a03a21){12b032(3b03a02)p031+3a02a03a21p031a03b03[9a21p031+2a02p1216b03(2p021+p121)]a032[4a21p021+(3b03a02)q121]},\begin{array}[]{ll}d_{11}=&\tfrac{1}{2(a_{02}-3b_{03})(3b_{03}^{2}-a_{03}a_{21})}\big{\{}a_{02}a_{21}p_{031}-b_{03}\big{[}3a_{21}p_{031}-6b_{03}(p_{021}-p_{121})-2a_{02}p_{121}\big{]}\\[2.15277pt] &-a_{03}\big{[}2a_{21}p_{021}+(a_{02}-3b_{03})q_{121}\big{]}\big{\}},\\[0.0pt] d_{21}=&0,\\ d_{31}=&\tfrac{1}{6a_{03}(3b_{03}^{2}-a_{03}a_{21})}(3a_{21}b_{03}p_{031}+2a_{03}a_{21}p_{121}-3a_{03}b_{03}q_{121}),\\[4.30554pt] d_{41}=&\tfrac{1}{a_{03}(3b_{03}-a_{02})}(a_{03}p_{021}-a_{02}p_{031}+3b_{03}p_{031}),\\[4.30554pt] d_{51}=&\tfrac{1}{2a_{03}(a_{02}-3b_{03})(3b_{03}^{2}-a_{03}a_{21})}\big{\{}12b_{03}^{2}(3b_{03}-a_{02})p_{031}+3a_{02}a_{03}a_{21}p_{031}\\[2.15277pt] &-a_{03}b_{03}\big{[}9a_{21}p_{031}+2a_{02}p_{121}-6b_{03}(2p_{021}+p_{121})\big{]}-a_{03}^{2}\big{[}4a_{21}p_{021}+(3b_{03}-a_{02})q_{121}\big{]}\big{\}},\end{array}

eliminates the terms y2y^{2}, xy2xy^{2}, y3y^{3} in the x˙\dot{x} equation, and the terms yy, xy2xy^{2} in the y˙\dot{y} equation. It implies that the perturbations

k=1εk(p02ky2+p03ky3+p12kxy2)andk=1εkq12kxy2,\sum_{k=1}\varepsilon^{k}(p_{02k}y^{2}+p_{03k}y^{3}+p_{12k}xy^{2})\qquad\text{and}\qquad\sum_{k=1}\varepsilon^{k}q_{12k}xy^{2}, (50)

in (45) are redundant and can be removed. Thus, we assume that p02k=p03k=p12k=q12k=0p_{02k}=p_{03k}=p_{12k}=q_{12k}=0. Further, introducing the transformation, (x,y,t)(ε3x,ε2y,tε)(x,y,t)\rightarrow(\varepsilon^{3}x,\varepsilon^{2}y,\frac{t}{\varepsilon}) into system (45) yields

(x˙y˙)={(δ1xy+12δ1(3ε3x2+ε6x3)+2(a21+k=1p21kεk)εxy+a03ε2y3+(a023b03)y2+(a21+k=1p21kεk)ε4x2y3b03ε3xy2,(1b2ε3)x+δ1y+12(3ε3bε6)x2+2k=1q21kεk+2xy(a21k=1q02kεk)εy2+12ε6x3+k=1q21kεk+5x2ya21ε4xy2+(b03+k=1q03kεk)ε3y3,),ify>0,(δ1xy+12δ1(3ε3x2+ε6x3)+2(a21+k=1p21kεk)εxy+a03ε2y3(a02+3b03)y2+(a21+k=1p21kεk)ε4x2y3b03ε3xy2,b+(1+32bε3)x+δ1y+12(3ε3+bε6)x2+2k=1q21kεk+2xy(a21+k=1q02kε2)εy2+12ε6x3+k=1q21kεk+5x2ya21ε4xy2+(b03+k=1q03kεk)ε3y3,),ify<0.{\small\left(\begin{array}[]{cc}\dot{x}\\ \dot{y}\end{array}\right)=\left\{\begin{aligned} &\left(\begin{aligned} &\delta_{1}x-y+\tfrac{1}{2}\delta_{1}(3\varepsilon^{3}x^{2}+\varepsilon^{6}x^{3})+2(a_{21}+\sum_{k=1}p_{21k}\varepsilon^{k})\varepsilon xy+a_{03}\varepsilon^{2}y^{3}\\ &+(a_{02}-3b_{03})y^{2}+(a_{21}+\sum_{k=1}p_{21k}\varepsilon^{k})\varepsilon^{4}x^{2}y-3b_{03}\varepsilon^{3}xy^{2},\\[1.93748pt] &(1-\tfrac{b}{2}\varepsilon^{3})x+\delta_{1}y+\tfrac{1}{2}(3\varepsilon^{3}-b\varepsilon^{6})x^{2}+2\sum_{k=1}q_{21k}\varepsilon^{k+2}xy\\ &-(a_{21}-\sum_{k=1}q_{02k}\varepsilon^{k})\varepsilon y^{2}+\tfrac{1}{2}\varepsilon^{6}x^{3}+\sum_{k=1}q_{21k}\varepsilon^{k+5}x^{2}y\\ &-a_{21}\varepsilon^{4}xy^{2}+(b_{03}+\sum_{k=1}q_{03k}\varepsilon^{k})\varepsilon^{3}y^{3},\\ \end{aligned}\right)\!,&\text{if}\ y>0,\\[3.87498pt] &\left(\begin{aligned} &\delta_{1}x-y+\tfrac{1}{2}\delta_{1}(3\varepsilon^{3}x^{2}+\varepsilon^{6}x^{3})+2(a_{21}+\sum_{k=1}p_{21k}\varepsilon^{k})\varepsilon xy+a_{03}\varepsilon^{2}y^{3}\\ &-(a_{02}+3b_{03})y^{2}+(a_{21}+\sum_{k=1}p_{21k}\varepsilon^{k})\varepsilon^{4}x^{2}y-3b_{03}\varepsilon^{3}xy^{2},\\[1.93748pt] &b+(1+\tfrac{3}{2}b\varepsilon^{3})x+\delta_{1}y+\tfrac{1}{2}(3\varepsilon^{3}+b\varepsilon^{6})x^{2}+2\sum_{k=1}q_{21k}\varepsilon^{k+2}xy\\ &-(a_{21}+\sum_{k=1}q_{02k}\varepsilon^{2})\varepsilon y^{2}+\tfrac{1}{2}\varepsilon^{6}x^{3}+\sum_{k=1}q_{21k}\varepsilon^{k+5}x^{2}y\\ &-a_{21}\varepsilon^{4}xy^{2}+(b_{03}+\sum_{k=1}q_{03k}\varepsilon^{k})\varepsilon^{3}y^{3},\end{aligned}\right)\!,&\text{if}\ y<0.\end{aligned}\right.} (51)

To further simplify system (51), we apply the following scaling,

a21=ε2A21,a02=ε3A02,b03=ε3B03,a03=ε4A03,q03k=ε3kQ03k,p21k=ε2kP21k,q02k=ε2kQ02k,q21k=1εk1Q21k,\begin{array}[]{lllll}a_{21}=\varepsilon^{2}A_{21},&a_{02}=\varepsilon^{3}A_{02},&b_{03}=\varepsilon^{3}B_{03},&a_{03}=\varepsilon^{4}A_{03},\\[4.30554pt] q_{03k}=\varepsilon^{3-k}Q_{03k},&p_{21k}=\varepsilon^{2-k}P_{21k},&q_{02k}=\varepsilon^{2-k}Q_{02k},&q_{21k}=\dfrac{1}{\varepsilon^{k-1}}\,Q_{21k},&\end{array} (52)

to (51), and obtain

(x˙y˙)={(y+δ1x+12δ1(3ε3x2+ε6x3)+ε3[2(A21+k=1P21k)xy+(A023B03)y2]+ε6[(A21+k=1P21k)x2y3B03xy2+A03y3],xb2ε3x+δ1y12bε6x2+ε3[32x2+2k=1Q21kxy(A21k=1Q02k)y2]+ε6[12x3+k=1Q21kx2yA21xy2+(B03+k=1Q03k)y3],),ify>0,(y+δ1x+12δ1(3ε4x2+ε7x3)+ε3[2(A21+k=1P21k)xy(A02+3B03)y2]+ε6[(A21+k=1P21k)x2y3B03xy2+A03y3],x+b+3b2ε3x+δ1y+b2ε6x2+ε3[32x2+2k=1Q21kxy(A21+k=1Q02k)y2]+ε6[12x3+k=1Q21kx2yA21xy2+(B03+k=1Q03k)y3],),ify<0.{\small\left(\begin{array}[]{cc}\dot{x}\\ \dot{y}\end{array}\right)=\left\{\begin{aligned} &\left(\begin{aligned} &-y+\delta_{1}x+\tfrac{1}{2}\delta_{1}(3\varepsilon^{3}x^{2}+\varepsilon^{6}x^{3})+\varepsilon^{3}\big{[}2(A_{21}+\sum_{k=1}P_{21k})xy\\[1.93748pt] &+(A_{02}-3B_{03})y^{2}\big{]}+\varepsilon^{6}\big{[}(A_{21}+\sum_{k=1}P_{21k})x^{2}y-3B_{03}xy^{2}+A_{03}y^{3}\big{]},\\[1.93748pt] &x-\tfrac{b}{2}\varepsilon^{3}x+\delta_{1}y-\tfrac{1}{2}b\varepsilon^{6}x^{2}+\varepsilon^{3}\big{[}\tfrac{3}{2}x^{2}+2\sum_{k=1}Q_{21k}xy-(A_{21}-\sum_{k=1}Q_{02k})y^{2}\big{]}\\ &+\varepsilon^{6}\big{[}\tfrac{1}{2}x^{3}+\sum_{k=1}Q_{21k}x^{2}y-A_{21}xy^{2}+(B_{03}+\sum_{k=1}Q_{03k})y^{3}\big{]},\\ \end{aligned}\right)\!,&\text{if}\ y>0,\\[7.74997pt] &\left(\begin{aligned} &-y+\delta_{1}x+\tfrac{1}{2}\delta_{1}(3\varepsilon^{4}x^{2}+\varepsilon^{7}x^{3})+\varepsilon^{3}\big{[}2(A_{21}+\sum_{k=1}P_{21k})xy\\[1.93748pt] &-(A_{02}+3B_{03})y^{2}\big{]}+\varepsilon^{6}\big{[}(A_{21}+\sum_{k=1}P_{21k})x^{2}y-3B_{03}xy^{2}+A_{03}y^{3}\big{]},\\[1.93748pt] &x+b+\tfrac{3b}{2}\varepsilon^{3}x+\delta_{1}y+\tfrac{b}{2}\varepsilon^{6}x^{2}+\varepsilon^{3}\big{[}\tfrac{3}{2}x^{2}+2\sum_{k=1}Q_{21k}xy-(A_{21}\\ &+\sum_{k=1}Q_{02k})y^{2}\big{]}+\varepsilon^{6}\big{[}\tfrac{1}{2}x^{3}+\sum_{k=1}Q_{21k}x^{2}y-A_{21}xy^{2}+(B_{03}+\sum_{k=1}Q_{03k})y^{3}\big{]},\end{aligned}\right)\!,&\text{if}\ y<0.\end{aligned}\right.} (53)

Therefore, the singular point (1,0)(1,0) of system (9) corresponds to the origin of system (53). To compute the generalized Lyapunov constants Vk(ε)V_{k}(\varepsilon) (k2k\geq 2), we let b=δ1=0b=\delta_{1}=0, which yields V1(ε)=0V_{1}(\varepsilon)=0. It is seen from (53) that there exists an infinite number of parameters, but most of them are redundant. Without loss of generality, we let

P21k=Q21k=Q02k=Q03k=0,k1,k2,\begin{array}[]{lllll}P_{21k}=Q_{21k}\!\!\!\!&=Q_{02k}\!\!\!\!&=Q_{03k}=0,\ \ \forall k\geq 1,\ k\neq 2,\end{array} (54)

which, together with b=δ1=0b=\delta_{1}=0, are substituted into (53) to yield the final system,

(x˙y˙)={(y+ε3[2(A21+P212)xy+(A023B03)y2]+ε6[(A21+P212)x2y3B03xy2+A03y3],x+ε3[32x2+2Q212xy(A21Q022)y2]+ε6[12x3+Q212x2yA21xy2+(B03+Q032)y3],),ify>0,(y+ε3[2(A21+P212)xy(A02+3B03)y2]+ε6[(A21+P212)x2y3B03xy2+A03y3],x+ε3[32x2+2Q212xy(A21+Q022)y2]+ε6[12x3+Q212x2yA21xy2+(B03+Q032)y3],),ify<0,\left(\begin{array}[]{cc}\dot{x}\\ \dot{y}\end{array}\right)=\left\{\begin{aligned} &\left(\begin{aligned} &-y+\varepsilon^{3}\big{[}2(A_{21}+P_{212})xy+(A_{02}-3B_{03})y^{2}\big{]}\\[2.15277pt] &+\varepsilon^{6}\big{[}(A_{21}+P_{212})x^{2}y-3B_{03}xy^{2}+A_{03}y^{3}\big{]},\\[2.15277pt] &x+\varepsilon^{3}\big{[}\tfrac{3}{2}x^{2}+2Q_{212}xy-(A_{21}-Q_{022})y^{2}\big{]}\\ &+\varepsilon^{6}\big{[}\tfrac{1}{2}x^{3}+Q_{212}x^{2}y-A_{21}xy^{2}+(B_{03}+Q_{032})y^{3}\big{]},\\ \end{aligned}\right)\!,&\text{if}\ y>0,\\[4.30554pt] &\left(\begin{aligned} &-y+\varepsilon^{3}\big{[}2(A_{21}+P_{212})xy-(A_{02}+3B_{03})y^{2}\big{]}\\[2.15277pt] &+\varepsilon^{6}\big{[}(A_{21}+P_{212})x^{2}y-3B_{03}xy^{2}+A_{03}y^{3}\big{]},\\[2.15277pt] &x+\varepsilon^{3}\big{[}\tfrac{3}{2}x^{2}+2Q_{212}xy-(A_{21}+Q_{022})y^{2}\big{]}\\ &+\varepsilon^{6}\big{[}\tfrac{1}{2}x^{3}+Q_{212}x^{2}y-A_{21}xy^{2}+(B_{03}+Q_{032})y^{3}\big{]},\end{aligned}\right)\!,&\text{if}\ y<0,\end{aligned}\right. (55)

for computing the generalized Lyapunov constants Vk(ε)(k2)V_{k}(\varepsilon)\ (k\geq 2). Returning to system (51), (55) becomes

(x˙y˙)={(y+δ1(x+32ε3x2+12ε6x3)+2ε(a21+ε2p212)xy+(a023b03)y2+ε4(a21+ε2p212)x2y3ε3b03xy2+ε2a03y3,xb2ε3(x+ε3x2)+δ1y+32ε3x2+2ε4q212xyε(a21ε2q022)y2+12ε6x3+ε7q212x2yε4a21xy2+ε3(b03+ε2q032)y3,),ify>0,(y+δ1(x+32ε3x2+12ε6x3)+2ε(a21+ε2p212)xy(a02+3b03)y2+ε4(a21+ε2p212)x2y3ε3b03xy2+ε2a03y3,x+b+b2ε3(3x+ε3x2)+δ1y+32ε3x2+2ε4q212xyε(a21+ε2q022)y2+12ε6x3+ε7q212x2yε4a21xy2+ε3(b03+ε2q032)y3,),ify<0.{\small\left(\begin{array}[]{cc}\dot{x}\\ \dot{y}\end{array}\right)=\left\{\begin{aligned} &\left(\begin{aligned} &-y+\delta_{1}\big{(}x+\tfrac{3}{2}\varepsilon^{3}x^{2}+\tfrac{1}{2}\varepsilon^{6}x^{3}\big{)}+2\,\varepsilon\,(a_{21}+\varepsilon^{2}p_{212})xy\\ &+(a_{02}-3b_{03})y^{2}+\varepsilon^{4}(a_{21}+\varepsilon^{2}p_{212})x^{2}y-3\,\varepsilon^{3}b_{03}xy^{2}+\varepsilon^{2}a_{03}y^{3},\\[1.93748pt] &x-\tfrac{b}{2}\varepsilon^{3}(x+\varepsilon^{3}x^{2})+\delta_{1}y+\tfrac{3}{2}\varepsilon^{3}x^{2}+2\varepsilon^{4}q_{212}xy-\varepsilon(a_{21}-\varepsilon^{2}q_{022})y^{2}\\ &+\tfrac{1}{2}\varepsilon^{6}x^{3}+\varepsilon^{7}q_{212}x^{2}y-\varepsilon^{4}a_{21}xy^{2}+\varepsilon^{3}(b_{03}+\varepsilon^{2}q_{032})y^{3},\\ \end{aligned}\right)\!,&\text{if}\ y>0,\\[3.87498pt] &\left(\begin{aligned} &-y+\delta_{1}\big{(}x+\tfrac{3}{2}\varepsilon^{3}x^{2}+\tfrac{1}{2}\varepsilon^{6}x^{3}\big{)}+2\,\varepsilon(a_{21}+\varepsilon^{2}p_{212})xy\\ &-(a_{02}+3b_{03})y^{2}+\varepsilon^{4}(a_{21}+\varepsilon^{2}p_{212})x^{2}y-3\varepsilon^{3}b_{03}xy^{2}+\varepsilon^{2}a_{03}y^{3},\\[1.93748pt] &x+b\!+\!\tfrac{b}{2}\varepsilon^{3}(3x\!+\!\varepsilon^{3}x^{2})\!+\!\delta_{1}y\!+\!\tfrac{3}{2}\varepsilon^{3}x^{2}\!+\!2\varepsilon^{4}q_{212}xy-\varepsilon(a_{21}\!+\!\varepsilon^{2}q_{022})y^{2}\\ &+\tfrac{1}{2}\varepsilon^{6}x^{3}+\varepsilon^{7}q_{212}x^{2}y-\varepsilon^{4}a_{21}xy^{2}+\varepsilon^{3}(b_{03}+\varepsilon^{2}q_{032})y^{3},\end{aligned}\right)\!,&\text{if}\ y<0.\end{aligned}\right.} (56)

We will show that system (56) can have at least 1010 small-amplitude limit cycles around the origin.

Now, system (55) has only 88 independent parameters: Q022Q_{022}, Q032Q_{032}, Q212Q_{212}, A03A_{03}, A21A_{21}, B03B_{03}, A02A_{02} and P212P_{212}, which will be used to solve the Lyapunov equations Vk(ε)=0V_{k}(\varepsilon)=0. Note that later we will perturb the constant term bb to obtain one more limit cycle by using the pseudo-Hopf bifurcation. The 1st generalized Lyapunov constant for system (55) is given by V1(ε)=2πδ1=0V_{1}(\varepsilon)=2\pi\delta_{1}=0 when δ1=0\delta_{1}=0. Then, we have the 2nd generalized Lyapunov constant,

V2(ε)=\displaystyle V_{2}(\varepsilon)= 83ε3Q022.\displaystyle\frac{8}{3}\varepsilon^{3}\,Q_{022}.

Setting Q022=0Q_{022}=0 yields V2(ε)=0V_{2}(\varepsilon)=0 and the 3rd generalized Lyapunov constant,

V3(ε)=\displaystyle V_{3}(\varepsilon)= 14πε6[3Q0322(1A21)Q2126B03P212].\displaystyle\frac{1}{4}\pi\varepsilon^{6}\big{[}3Q_{032}-2(1-A_{21})Q_{212}-6B_{03}P_{212}\big{]}.

We obtain

Q032=23(1A21)Q212+2B03P122\displaystyle Q_{032}=\frac{2}{3}(1-A_{21})\,Q_{212}+2B_{03}P_{122}

by solving V3(ε)=0V_{3}(\varepsilon)=0. Further, we have the 4th generalized Lyapunov constant,

V4(ε)=\displaystyle V_{4}(\varepsilon)= 1645ε9A02[(3+P212)Q21212B03P212].\displaystyle\frac{16}{45}\varepsilon^{9}A_{02}\big{[}(3+P_{212})Q_{212}-12B_{03}P_{212}\big{]}.

Since the condition required for the center condition VI does not allow a02=0a_{02}=0, we have A020A_{02}\neq 0. Thus, solving V4(ε)=0V_{4}(\varepsilon)=0 for Q212Q_{212} we obtain

Q212=\displaystyle Q_{212}= 12B03P2123+2P212,(3+2P2120).\displaystyle\dfrac{12B_{03}P_{212}}{3+2P_{212}},\quad(3+2P_{212}\neq 0).

Then, solving the 5th generalized Lyapunov constant equation V5(ε)=0V_{5}(\varepsilon)=0 for A03A_{03} yields

A03=\displaystyle A_{03}= 13(2P2121)(3+2P212)2{(3+2P212)2[15A022+2(A21+P212+1)(A21(10P212+3)\displaystyle\dfrac{1}{3(2P_{212}-1)(3+2P_{212})^{2}}\big{\{}(3+2P_{212})^{2}\big{[}15A_{02}^{2}+2(A_{21}+P_{212}+1)\big{(}A_{21}(10P_{212}+3)
+4P2122+4P212+9)]72B032(6P212+5)(8A21P212+2P2122P212+6)},\displaystyle+4P_{212}^{2}+4P_{212}+9\big{)}\big{]}-72B_{03}^{2}(6P_{212}+5)\big{(}8A_{21}P_{212}+2P_{212}^{2}-P_{212}+6\big{)}\big{\}},

leading to

V6(ε)=\displaystyle V_{6}(\varepsilon)= 128ε15A02B03P212525(2P2121)(3+2P212)3V6a,\displaystyle-\frac{128\varepsilon^{15}A_{02}B_{03}P_{212}}{525(2P_{212}-1)(3+2P_{212})^{3}}\,V_{6a},
V7(ε)=\displaystyle V_{7}(\varepsilon)= ε18B03P21250400(2P2121)2(3+2P212)5V7a,\displaystyle-\frac{\varepsilon^{18}B_{03}P_{212}}{50400(2P_{212}-1)^{2}(3+2P_{212})^{5}}\,V_{7a},
V8(ε)=\displaystyle V_{8}(\varepsilon)= ε21B03P2126350400(2P2121)2(3+2P212)5V8a,\displaystyle-\frac{\varepsilon^{21}B_{03}P_{212}}{6350400(2P_{212}-1)^{2}(3+2P_{212})^{5}}\,V_{8a},
V9(ε)=\displaystyle V_{9}(\varepsilon)= ε24B03P21276204800(2P2121)3(3+2P212)7V9a,\displaystyle-\frac{\varepsilon^{24}B_{03}P_{212}}{76204800(2P_{212}-1)^{3}(3+2P_{212})^{7}}\,V_{9a},
V10(ε)=\displaystyle V_{10}(\varepsilon)= ε27B03P21276204800(2P2121)3(3+2P212)7V10a,\displaystyle-\frac{\varepsilon^{27}B_{03}P_{212}}{76204800(2P_{212}-1)^{3}(3+2P_{212})^{7}}\,V_{10a},

where V6aV_{6a}, V7aV_{7a}, V8aV_{8a}, V9aV_{9a}, and V10aV_{10a} are polynomials in A21A_{21}, B032B_{03}^{2}, A02A_{02} and P212P_{212}. In particular,

V6a=\displaystyle V_{6a}= 216B032[24A21P212(4P212220P21215)+16P2124224P2123280P212265]\displaystyle\ 216B_{03}^{2}\big{[}24A_{21}P_{212}\big{(}4P_{212}^{2}-20P_{212}-15\big{)}+16P_{212}^{4}-224P_{212}^{3}-280P_{212}-265\big{]}
+(3+2P212)2[405A022+A21(864A21P212+1148P2122+728P212+645)\displaystyle+(3+2P_{212})^{2}\big{[}405A_{02}^{2}+A_{21}\big{(}864A_{21}P_{212}+1148P_{212}^{2}+728P_{212}+645\big{)}
+(P212+1)(284P2122136P212+645)],\displaystyle+(P_{212}+1)(284P_{212}^{2}-136P_{212}+645)\big{]},

and other lengthy expressions of V7aV_{7a}, V8aV_{8a}, V9aV_{9a}, and V10aV_{10a} are omitted for brevity. Here, we assume that

(2P2121)(3+2P212)0.(2P_{212}-1)(3+2P_{212})\neq 0. (57)

Then, we use Maple with its built-in command eliminate,

eliminate({V6a,V7a,V8a,V9a},A21),{\rm eliminate}(\{V_{6a},V_{7a},V_{8a},V_{9a}\},A_{21}),

to obtain a solution A21=A21nA21dA_{21}=-\,\frac{A_{21n}}{A_{21d}}, where

A21n=\displaystyle A_{21n}= 2539579392B036P2122(16P2124224P2123280P212265)\displaystyle\ 2539579392B_{03}^{6}P_{212}^{2}(16P_{212}^{4}-224P_{212}^{3}-280P_{212}-265)
+279936B034P212[17010A022P212(3+2P212)2+37408P2126+185392P2125\displaystyle+279936B_{03}^{4}P_{212}\big{[}17010A_{02}^{2}P_{212}(3+2P_{212})^{2}+37408P_{212}^{6}+185392P_{212}^{5}
+265584P2124+266792P2123+424802P2122+207315P212130248]\displaystyle+265584P_{212}^{4}+266792P_{212}^{3}+424802P_{212}^{2}+207315P_{212}-130248\big{]}
+54B032(3+2P212)2[243A022P212(20932P212226676P212+13977)\displaystyle+54B_{03}^{2}(3+2P_{212})^{2}\big{[}243A_{02}^{2}P_{212}\big{(}20932P_{212}^{2}-26676P_{212}+13977\big{)}
322112P2126+853408P2125+4354704P2124+4303208P2123+1241408P2122\displaystyle-322112P_{212}^{6}+853408P_{212}^{5}+4354704P_{212}^{4}+4303208P_{212}^{3}+1241408P_{212}^{2}
18636P2129540]+81A022(3+2P212)4(46388P2122+89553P212+45)\displaystyle-18636P_{212}-9540\big{]}+81A_{02}^{2}(3+2P_{212})^{4}(46388P_{212}^{2}+89553P_{212}+45)
5(3+2P212)4(2P2121)(P212+1)(7P212+3)(34P212+3)(58P212+129)\displaystyle-5(3+2P_{212})^{4}(2P_{212}-1)(P_{212}+1)(7P_{212}+3)(34P_{212}+3)(58P_{212}+129)

and

A21d=\displaystyle A_{21d}= 60949905408B036P2123(4P212220P21215)\displaystyle\ 60949905408B_{03}^{6}P_{212}^{3}(4P_{212}^{2}-20P_{212}-15)
+5038848B034P2122(3+2P212)2(1484P2122+3288P2121039)\displaystyle+5038848B_{03}^{4}P_{212}^{2}(3+2P_{212})^{2}(1484P_{212}^{2}+3288P_{212}-1039)
+3888B032P212(3+2P212)2(122472A022P2123416P2124+13700P2123\displaystyle+3888B_{03}^{2}P_{212}(3+2P_{212})^{2}(122472A_{02}^{2}P_{212}-3416P_{212}^{4}+13700P_{212}^{3}
+34622P2122+18267P212+1512)+118098A022P212(3+2P212)4\displaystyle+34622P_{212}^{2}+18267P_{212}+1512)+118098A_{02}^{2}P_{212}(3+2P_{212})^{4}
5(3+2P212)4(2P2121)(7P212+3)(34P212+3)(58P212+129),\displaystyle-5(3+2P_{212})^{4}(2P_{212}-1)(7P_{212}+3)(34P_{212}+3)(58P_{212}+129),

and three polynomial resultant equations,

Rk=Rk(B032,A02,P212)=0,k=1,2,3.R_{k}=R_{k}(B_{03}^{2},A_{02},P_{212})=0,\quad k=1,2,3.

Further, we apply the command eliminate again to the three resultants to obtain a solution B032=B032(A02,P212)B_{03}^{2}=B_{03}^{2}(A_{02},P_{212}), and two lengthy polynomial resultants,

R12=A02P212(3+2P212)R12aR12b,R13=A02P212(3+2P212)R13aR13b,\begin{array}[]{ll}R_{12}=A_{02}\,P_{212}(3+2P_{212})R_{12a}R_{12b},\quad R_{13}=A_{02}\,P_{212}(3+2P_{212})R_{13a}R_{13b},\end{array}

where R12aR_{12a}, R12bR_{12b}, R13aR_{13a} and R13bR_{13b} are lengthy polynomials in A02A_{02} and P212P_{212}, with 245245, 58805880, 30863086 and 1949819498 terms, respectively. There are four combinations: (R12a,R13a)(R_{12a},R_{13a}), (R12a,R13b)(R_{12a},R_{13b}), (R12b,R13a)(R_{12b},R_{13a}) and (R12b,R13b)(R_{12b},R_{13b}) to find the solutions for (A02,P212)(A_{02},P_{212}). However, it can be shown that only the combination (R12a,R13b)(R_{12a},R_{13b}) generates maximal number of limit cycles. Finally, we use the Maple-built command resultant,

resultant(R12a,R13a,A02),{\rm resultant}(R_{12a},R_{13a},A_{02}),

to obtain the following resultant,

R1213=\displaystyle\!\!\!\!R_{1213}= C~π14P21270(2P2121)20(3+2P212)36\displaystyle\ \tilde{C}\pi^{14}P_{212}^{70}(2P_{212}-1)^{20}(3+2P_{212})^{36} (58)
×[(P212+2)6(18000P2124+111504P2123+182360P2122+98444P212+9965)\displaystyle\times\big{[}(P_{212}+2)^{6}(18000P_{212}^{4}+111504P_{212}^{3}+182360P_{212}^{2}+98444P_{212}+9965)
×(33888P2126+237600P2125+463840P2124+311472P2123\displaystyle\quad\ \times\!(33888P_{212}^{6}+237600P_{212}^{5}+463840P_{212}^{4}+311472P_{212}^{3}
+75078P2122+7338P212+243)]R1213aR1213b,\displaystyle\qquad\ +\!75078P_{212}^{2}+7338P_{212}+243)\big{]}\,R_{1213a}R_{1213b},

where C~\tilde{C} is a big positive integer, R1213aR_{1213a} and R1213bR_{1213b} are respectively 5454th-degree and 10801080th-degree polynomials in P212P_{212}. With the aid of Maple using 10001000 digits accuracy, we obtain 99, 2222 and 221221 real solutions from the terms in the script bracket of R1213R_{1213}, the polynomials R1213aR_{1213a} and R1213bR_{1213b}, respectively.

Remark 5.1.

In most research articles, researchers prefer to not use numerical computation in a rigorous mathematical proof. Here, we should point out that (1) our numerical computation is only used at the last step, and all computations in the previous steps are symbolic and exact. (2) we can, instead of using numerical computation, apply the interval computation (or interval arithmetic) [27] to prove the existence of the solutions in an interval satisfying a required accuracy. This does not really change the basics of computation, and a numerical computation makes the presentation simple and clear. (3) the most important point is that the accuracy to be taken (e.g., 10001000 decimal points) should be good enough to guarantee the reliability of proof. That is, the accumulation of round errors in the numerical computation does not affect the conclusion for the existence of solutions. In other words, it is not a matter of numerical computation, but the high enough accuracy.

It can be shown that all the 221221 solutions obtained from R1213bR_{1213b} produces V1=V2==V8=0V_{1}=V_{2}=\cdots=V_{8}=0, but V90V_{9}\neq 0, giving 8×2=168\times 2=16 small amplitude limit cycles. While from the 9+229+22 solutions, we obtain 88 solutions (with δ1=Q022=0\delta_{1}=Q_{022}=0):

(P212,A02,B03,A21,A03,Q212,Q032)=(84.39736368, 246.66126379,±2.18680584,30.37184157,1169.84368584,±13.35825220,630.69227526),(2.12093549,0.68942957,±0.05796919,1.33657024,0.13502593,±1.18803564,0.51246946),(1.64608697,2.82165606,±0.10278968,20.81805724,0.15805791,±6.94931604,100.74198202),(0.95226778,0.20225860,±0.06175799,0.20115348,0.09330998,0.64422154,0.63349293),\begin{array}[]{rlrlrlrl}&(P_{212},\,A_{02},\,B_{03},\,A_{21},\,A_{03},\,Q_{212},\,Q_{032})&&&\\[2.15277pt] =\!\!\!\!&(\hskip 10.47945pt-84.39736368\cdots,\,246.66126379\cdots,\hskip 11.70792pt\pm 2.18680584\cdots,\hskip 7.22743pt30.37184157\cdots,\\ &\ -1169.84368584\cdots,\pm 13.35825220\cdots,\mp 630.69227526\cdots),\\ &(\hskip 15.89948pt-2.12093549\cdots,\hskip 14.02039pt0.68942957\cdots,\hskip 10.98451pt\pm 0.05796919\cdots,\hskip 12.64746pt1.33657024\cdots,\\ &\hskip 28.90755pt0.13502593\cdots,\hskip 5.20389pt\pm 1.18803564\cdots,\hskip 10.98451pt\mp 0.51246946\cdots),\\ &(\hskip 15.89948pt-1.64608697\cdots,\hskip 14.02039pt2.82165606\cdots,\hskip 10.98451pt\pm 0.10278968\cdots,-20.81805724\cdots,\\ &\hskip 20.09105pt-0.15805791\cdots,\hskip 6.14343pt\pm 6.94931604\cdots,\mp 100.74198202\cdots),\\ &(\hskip 15.89948pt-0.95226778\cdots,\hskip 14.02039pt0.20225860\cdots,\hskip 10.98451pt\pm 0.06175799\cdots,\hskip 5.78172pt-0.20115348\cdots,\\ &\hskip 28.90755pt0.09330998\cdots,\hskip 6.14343pt\mp 0.64422154\cdots,\hskip 10.98451pt\mp 0.63349293\cdots),\\ \end{array}

which result in V2=V3==V9=0V_{2}=V_{3}=\cdots=V_{9}=0, but V100V_{10}\neq 0. Moreover, all the 88 solutions satisfy the requirement of the center condition VI,

a02+|a12|=a02+|3b03|=A02ε3+|3B03ε3|<0A02|3B03|>0(ε<0).a_{02}+|a_{12}|=a_{02}+|-3b_{03}|=A_{02}\varepsilon^{3}+|3B_{03}\varepsilon^{3}|<0\ \ \Longleftrightarrow\ \ A_{02}-|3B_{03}|>0\ \ (\varepsilon<0).

We choose one of the 88 solutions,

(P212,A02,B03,A21,A03,Q212,Q032,Q022,δ1)C=(2.12093549, 0.68942957,0.05796919, 1.33657024,0.13502593, 1.18803564,0.51246946, 0.0,0.0),\begin{array}[]{rlrlrlrl}&(P_{212},\,A_{02},\,B_{03},\,A_{21},\,A_{03},\,Q_{212},\,Q_{032},\,Q_{022},\,\delta_{1})_{\rm C}&&&\\[4.30554pt] =\!\!\!\!&(-2.12093549\cdots,\ 0.68942957\cdots,\hskip 10.84006pt0.05796919\cdots,\ 1.33657024\cdots,\\ &\hskip 13.00806pt0.13502593\cdots,\ 1.18803564\cdots,\,-0.51246946\cdots,\ 0.0,\hskip 57.81621pt0.0),\\ \end{array}

where the subscript C indicates a critical point, under which

V1=V2=V3==V9=0,V10=0.05995729ε27<0(ε<0).V_{1}=V_{2}=V_{3}=\cdots=V_{9}=0,\quad V_{10}=0.05995729\cdots\varepsilon^{27}<0\ \ (\varepsilon<0).

More precisely, with the accuracy of 10001000 decimal points, we have

V1=0.0,V2=0.0ε3,V3=0.0ε6,V4=0.63836071×10999ε9,V5= 0.26153383×10997ε12,V6=0.62071107×10970ε15,V7= 0.32934572×10969ε18,V8=0.10608263×10968ε21,V9= 0.27441458×10968ε24,\begin{array}[]{rlll}V_{1}=\!\!\!\!&0.0\,,\quad V_{2}=0.0\,\varepsilon^{3},\quad V_{3}=0.0\,\varepsilon^{6},\\[2.15277pt] V_{4}=\!\!\!\!&0.63836071\cdots\times 10^{-999}\,\varepsilon^{9},\hskip 14.45377ptV_{5}=\ \ \ 0.26153383\cdots\times 10^{-997}\,\varepsilon^{12},\\[2.15277pt] V_{6}=\!\!\!\!&0.62071107\cdots\times 10^{-970}\,\varepsilon^{15},\quad V_{7}=-\,0.32934572\cdots\times 10^{-969}\,\varepsilon^{18},\\[2.15277pt] V_{8}=\!\!\!\!&0.10608263\cdots\times 10^{-968}\,\varepsilon^{21},\quad V_{9}=-\,0.27441458\cdots\times 10^{-968}\,\varepsilon^{24},\end{array}

for which we are definitely confident that a solution exists, satisfying V1=V2==V9=0V_{1}=V_{2}=\cdots=V_{9}=0. In addition, a direct calculation shows that

det[(V1(ε),V2(ε),V3(ε),V4(ε),V5(ε),V6(ε),V7(ε),V8(ε),V9(ε))(δ1,Q022,Q032,Q212,A03,A21,B03,A02,P212)]C=329.04383261ε108>0.\det\left[\frac{\partial(V_{1}(\varepsilon),V_{2}(\varepsilon),V_{3}(\varepsilon),V_{4}(\varepsilon),V_{5}(\varepsilon),V_{6}(\varepsilon),V_{7}(\varepsilon),V_{8}(\varepsilon),V_{9}(\varepsilon))}{\partial(\delta_{1},Q_{022},Q_{032},Q_{212},A_{03},A_{21},B_{03},A_{02},P_{212})}\right]_{\rm C}\!\!\!=329.04383261\cdots\,\varepsilon^{108}>0.

Hence, by Lemma 3.1 we know that system (53) has at least 99 small-amplitude limit cycles bifurcating from the origin, leading to the existence of 1818 small-amplitude limit cycles bifurcating from the bi-center of such Z2Z_{2}-equivariant cubic switching systems.

Further, applying the pseudo-Hopf bifurcation, we obtain one more small-amplitude limit cycle for system (56) by perturbing the coefficient bb. For bb and ε\varepsilon sufficiently small, the switching system (56) has a small sliding segment on the switching manifold y=0y=0 with the end points at (0,0)(0,0) and (ϱb,0)(\varrho_{b},0), where ϱb\varrho_{b} is the root of the equation g(x,0)=0g^{-}(x,0)=0 in (56), namely,

x+b+b2ε3(3x+ε3x2)+32ε3x2+12ε6x3=12(b+x)(1+ε3x)(2+ε3x)=0,x+b+\dfrac{b}{2}\varepsilon^{3}(3x+\varepsilon^{3}x^{2})+\dfrac{3}{2}\varepsilon^{3}x^{2}+\dfrac{1}{2}\varepsilon^{6}x^{3}=\dfrac{1}{2}(b+x)(1+\varepsilon^{3}x)(2+\varepsilon^{3}x)=0,

which yields the solution ϱb=b\varrho_{b}=-b. Thus, the sliding segment shrinks to (0,0)(0,0) when bb goes to zero.

We consider the point (ϱε,0)(\varrho_{\varepsilon},0) on the switching line y=0y=0 with sufficiently small ϱb<ϱε\varrho_{b}<\varrho_{\varepsilon}, where ϱε\varrho_{\varepsilon} is defined as a bifurcation function, satisfying

d(ϱε,ε)=Υ+(ϱε,ε)(Υ)1(ϱε,ε).d(\varrho_{\varepsilon},\varepsilon)=\Upsilon^{+}(\varrho_{\varepsilon},\varepsilon)-(\Upsilon^{-})^{-1}(\varrho_{\varepsilon},\varepsilon). (59)

Then, we have two half-return maps given in the form of

Υ+(ϱε,ε)=V1+(ε)ϱε+O(ϱε2)and(Υ)1(ϱε,ε)=V0(ε)+V1(ε)ϱε+O(ϱε2),\Upsilon^{+}(\varrho_{\varepsilon},\varepsilon)=V_{1}^{+}(\varepsilon)\varrho_{\varepsilon}+O(\varrho_{\varepsilon}^{2})\quad\text{and}\quad(\Upsilon^{-})^{-1}(\varrho_{\varepsilon},\varepsilon)=V_{0}^{-}(\varepsilon)+V_{1}^{-}(\varepsilon)\varrho_{\varepsilon}+O(\varrho_{\varepsilon}^{2}), (60)

respectively, where

V0(ε)=b[2+O(ε3)]andV1±(ε)=±δ1π+O(ε3).V_{0}^{-}(\varepsilon)=b\,[2+O(\varepsilon^{3})]\quad\text{and}\quad V_{1}^{\pm}(\varepsilon)=\pm\delta_{1}\pi+O(\varepsilon^{3}). (61)

It follows from (59) and (60) that V0(ε)=b[2+O(ε3)]V_{0}(\varepsilon)=-b\,[2+O(\varepsilon^{3})] and V1(ε)|b=0=2δ1πV_{1}(\varepsilon)|_{b=0}=2\delta_{1}\pi, implying that V0(ε)=V1(ε)=0V_{0}(\varepsilon)=V_{1}(\varepsilon)=0 when b=δ1=0b=\delta_{1}=0. Then, to get more small-amplitude limit cycles, we let b=δ1=0b=\delta_{1}=0, and compute higher-order generalized Lyapunov constants. This means that we return to exactly where we start and obtain the system (55) under the assumption b=δ1=0b=\delta_{1}=0. Therefore, by Lemma 3.1 and combining the 99 small-amplitude limit cycles obtained above, we know that 1010 small-amplitude crossing limit cycles exist near the origin of system (56), and so 2020 small-amplitude limit cycles can bifurcate from the bi-center of the Z2Z_{2}-equivariant cubic switching system (9).

Note that since we include all εk\varepsilon^{k}-order perturbations in constructing system (41), the number 2020 is actually the maximal number of small-amplitude limit cycles which can be obtained with the 66 center conditions I-VI under Hopf and pseudo-Hopf bifurcations for the Z2Z_{2}-equivariant cubic switching system considered in this paper.

This finishes the proof for Theorem 2.3.

6 Conclusion

In this paper, we have studied the bi-center problem and the cyclicity problem for planar switching nilpotent systems. First, we generalize the Poincaré-Lyapunov method for switching systems with linear-type centers to compute the generalized Lyapunov constants for switching nilpotent systems. Then, we derive 66 bi-center conditions for cubic Z2Z_{2}-equivariant switching systems with two symmetric nilpotent singular points. In particular, we find that the bi-center (±1,0)(\pm 1,0) consists of the combination of two second-order nilpotent cusps. Further, we construct perturbed systems with the 66 center conditions and present one with the center condition VI to show the existence at least 2020 small-amplitude limit cycles bifurcating from the nilpotent bi-center, which provides a great improvement from 1212, leading to a new lower bound on the number of limit cycles in Z2Z_{2}-equivariant cubic switching systems.

Motivated from this work, some questions naturally arise, which may promote future research in this direction.

  1. 1.

    In this work, we use some special perturbation to obtain 66 center conditions for Z2Z_{2}-equivariant cubic switching system with nilpotent singular points. Does this special perturbation generate all possible center conditions? If not, what center conditions might be missed, and what kind of perturbations should be used to find all possible center conditions?

  2. 2.

    The computation of the Lyapunov constants heavily depends upon a computer algebra system and the algorithms to be used. From this work, we have found that the current techniques for symbolic computation is good enough for computing the generalized Lyapunov constants. However, the current well-known approaches such as Groebner basis, regular chains, and Maple built-in programs eliminate, resultant, etc. seem still not powerful enough to solve the Lyapunov constant (polynomial) equations as well as many other such equations related to Hilbert’s 16th problem. So, how to develop a more efficient method (or algorithm, or program) for solving the system of multivariate polynomial equations is a very challenging task and needs further research.

  3. 3.

    In this paper, our generalized Lyapunov constants contain different εk\varepsilon^{k} order terms (before scaling), and we treat them in a consistent way by scaling. However, in the literature, many works treat such Lyapunov constants or focus values according to the εk\varepsilon^{k} orders, similar to the consideration of the 1st-order, 2nd-order, etc. Melnikov functions (for example, see [43]). Then, what are the advantages and disadvantages of these two perturbation methods, and for what systems one of the two approaches is better to be applied?

Acknowledgment

This work was supported by the National Natural Science Foundation of China Nos. 12071091 (T. Chen), 12071198 (F. Li) and 12371175 (Y. Tian), the Guangdong Basic and Applied Basic Research Foundation No. 2022A1515011827 (T. Chen), the Project funded by China Postdoctoral Science Foundation No. 2023M744329 (T. Chen) and the Natural Sciences and Engineering Research Council of Canada, No. R2686A02 (P. Yu).

References

  • [1] A. Andronov, A. Vitt, S. Khaikin, Theory of Oscillations, Pergamon Press, Oxford, 1966.
  • [2] A. Algaba, C. García, J. Giné, J. Llibre, The center problem for Z2Z_{2}-symmetric nilpotent vector fields, J. Math. Anal. Appl. 466 (2018) 183–198.
  • [3] S. Banerjee, G. Verghese, Nonlinear Phenomena in Power Electronics: Attractors, Bifurcations, Chaos, and Nonlinear Control, Wiley-IEEE Press, New York, 2001.
  • [4] J. Bastos, C. Buzzi, J. Llibre, D. Novaes, Melnikov analysis in nonsmooth differential systems with nonlinear switching manifold, J. Differential Equations 267 (2019) 3748–3767.
  • [5] M.D. Bernardo, P. Kowalczyk, A.B. Nordmark, Sliding bifurcations: a novel mechanism for the sudden onset of chaos in dry friction oscillators, Int. J. Bifurcation Chaos 13 (2003) 2935–2948.
  • [6] F. Braun, L. P. C. da Cruz, J. Torregrosa, On the number of limit cycles in piecewise planar quadratic differential systemss, (2023) preprint.
  • [7] A. Buica˘\rm{\breve{a}}, J. Llibre, O. Makarenkov, Asymptotic stability of periodic solutions for nonsmooth differential equations with application to the nonsmooth van der Pol oscillator, SIAM J. Math. Anal. 40 (2009) 2478–2495.
  • [8] J. Castillo, J. Llibre, F. Verduzco, The pseudo-Hopf bifurcation for planar discontinuous piecewise linear differential systems, Nonlinear Dynam. 90 (2017) 1829–1840.
  • [9] X. Chen, Z. Du, Limit cycles bifurcate from centers of discontinuous quadratic systems, Comput. Math. Appl. 59 (2010) 3836–3848.
  • [10] X. Chen, W. Zhang, Isochronicity of centers in switching Bautin system, J. Differential Equations 252 (2012) 2877–2899.
  • [11] T. Chen, L. Huang, P. Yu, Center condition and bifurcation of limit cycles for quadratic switching systems with a nilpotent equilibrium point, J. Differential Equations 303 (2021) 326–368.
  • [12] T. Chen, L. Huang, P. Yu, W. Huang, Bifurcation of limit cycles at infinity in piecewise polynomial systems, Nonlinear Anal.: Real World Appl. 41 (2018) 82–106.
  • [13] T. Chen, S. Li, J. Llibre, Z2Z_{2}-equivariant linear type bi-center cubic polynomial Hamiltonian vector fields, J. Differential Equations 269 (2020) 832–861.
  • [14] I. Colak, J. Llibre and C. Valls, Hamiltonian nilpotent centers of linear plus cubic homogeneous polynomial vector fields, Adv. Math. 259 (2014) 655–687.
  • [15] L. Cruz, D. Novaes, J. Torregrosa, New lower bound for the Hilbert number in piecewise quadratic differential systems, J. Differential Equations 266 (2019) 4170–4203.
  • [16] F. Dumortier, J. Llibre, J. Artés, Qualitative Theory of Planar Differential Systems, Universitext, Springer-Verlag, New York, 2006.
  • [17] R.D. Euzébio, J. Llibre, D.J. Tonon, Lower bounds for the number of limit cycles in a generalised Rayleigh-Liénard oscillator, Nonlinearity 35 (2022) 3883–3906.
  • [18] A.F. Filippov, Differential Equation with Discontinuous Right-Hand Sides, Kluwer Academic, Netherlands, 1988.
  • [19] E. Freire, E. Ponce, F. Torres, A general mechanism to generate three limit cycles in planar Filippov systems with two zones, Nonlinear Dynam. 78 (2014) 251–263.
  • [20] I. García, Cyclicity of some symmetric nilpotent centers, J. Differential Equations 260 (2016) 5356–5377.
  • [21] A. Gasull, J. Torregrosa, Center-focus problem for discontinuous planar differential equations, Int. J. Bifurcation Chaos 13 (2003) 1755–1765.
  • [22] H. Giacomini, J. Giné, J. Llibre, The problem of distinguishing between a center and a focus for nilpotent and degenerate analytic systems, J. Differential Equations 227 (2006) 406–426.
  • [23] L. Gouveia, J. Torregrosa, Local cyclicity in low degree planar piecewise polynomial vector fields, Nonlinear Anal.: Real World Appl. 60 (2021) 103278 1-19.
  • [24] L. Guo, P. Yu, Y. Chen, Bifurcation analysis on a class of Z2Z_{2}-equivariant cubic switching systems showing eighteen limit cycles, J. Differential Equations 266 (2019) 1221–1244.
  • [25] M. Han, P. Yu, Normal Forms, Melnikov Functions and Bifurcations of Limit Cycles, Springer-Verlag, New York, 2012.
  • [26] M. Han, W. Zhang, On Hopf bifurcation in non-smooth planar system, J. Differential Equations 248 (2010) 2399–2416.
  • [27] T. Hickey, Q. Ju, M. H. Van Emden, Interval arithmetic: From principles to implementation, J. ACM 48 (2001) 1038–1068.
  • [28] S. Huan, X. Yang, Existence of limit cycles in general planar piecewise linear systems of saddle-saddle dynamics, Nonlinear Anal. 92 (2013) 82–85.
  • [29] M. Kunze, T. Kupper, Qualitative bifurcation analysis of a non-smooth friction oscillator model, Math. Phys. 48 (1997) 87–101.
  • [30] R. Leine, H. Nijmeijer, Dynamics and Bifurcations of Non-Smooth Mechanical Systems, Lect. Notes Appl. Comput. Mech., vol.18, Springer-Verlag, Berlin, 2004.
  • [31] F. Li, H. Li, Y. Liu, New double bifurcation of nilpotent focus, Int. J. Bifurcation Chaos 31 (2021) 2150053.
  • [32] F. Li, Y. Liu, Y. Liu, P. Yu, Bi-center problem and bifurcation of limit cycles from nilpotent singular points in Z2Z_{2}-equivariant cubic vector fields, J. Differential Equations 265 (2018) 4965–4992.
  • [33] F. Li, Y. Liu, P. Yu, J. Wang, Complex integrability and linearizability of cubic Z2Z_{2}-equivariant systems with two 1:q resonant singular points, J. Differential Equations 300 (2021) 786–813.
  • [34] F. Li, Y. Jin, Y. Tian, P. Yu, Integrability and linearizability of cubic Z2Z_{2} systems with non-resonant singular points, J. Differential Equations 269 (2020) 9026–9049.
  • [35] F. Li, P. Yu, Y. Tian, Y. Liu, Center and isochronous center conditions for switching systems associated with elementary singular points, Commun. Nonlinear Sci. Numer. Simul. 28 (2015) 81–97.
  • [36] J. Li, Hilbert’s 16th problem and bifurcations of planar polynomial vector fields, Int. J. Bifurcation Chaos 3 (2013) 47–106.
  • [37] J. Li, Y. Liu, New results on the study of ZqZ_{q}-equivariant planar polynomial vector fields, Qual. Theo. Dyna. Syst. 9 (2010) 167–219.
  • [38] J. Li, Y. Tian, Y. Liu, Twenty-six crossing limit cycles around one singular point in a cubic switching system, Int. J. Bifurcation Chaos 32(10) (2022) 2250158 (7 pages).
  • [39] Y. Liu, F. Li, Double bifurcation of nilpotent focus, Int. J. Bifurcation Chaos 25(3) (2015) 1550036 (10 pages).
  • [40] Y. Lv, R. Yuan, Y. Pei, Dynamics in two nonsmooth predator-prey models with threshold harvesting, Nonlinear Dynam. 74 (2013) 107–132.
  • [41] N. Minorsky, Nonlinear Oscillations, Van Nostrand, New York, 1962.
  • [42] Y. Tian, P. Yu, Center conditions in a switching Bautin system, J. Differential Equations 259 (2015) 1203–1226.
  • [43] Y. Tian, P. Yu, Bifurcation of small limit cycles in cubic integrable systems using higher-order analysis, J. Differential Equations 264(9) (2018) 5950–5976.
  • [44] E. Stróżzyna, H. Żoła̧dek, The analytic normal for the nilpotent singularity, J. Differential Equations 179 (2012) 479–537.
  • [45] J. Yang, L. Zhao, The cyclicity of period annuli for a class of cubic Hamiltonian systems with nilpotent singular points, J. Differential Equations 263(9) (2017) 5554–5581.
  • [46] P. Yu, M. Han, X. Zhang, Eighteen limit cycles around two symmetric foci in a cubic planar switching polynomial system, J. Differential Equations 275(2) (2021) 939–959.